# HarperFlow — full site content
> Markdown of HarperFlow's product, research, comparison, guide and documentation pages, generated from www.harperflow.io. Index: https://www.harperflow.io/llms.txt
---
Source: https://www.harperflow.io/
# Turn your blog into an AI-search growth engine
HarperFlow runs your Webflow blog on autopilot—researching, writing, styling, and publishing source-backed articles built for AI search. Also available for WordPress and Shopify.
[](https://apps.shopify.com/harperflow) [](https://webflow.com/apps/detail/harperflow) [](https://wordpress.org/plugins/harperflow/)
## 8–12
cited sources in every article
## 117
AI answers analysed in our published research
## Outsmart the changing search landscape.
Growth not tied to ad spend
Every paid click resets the meter. HarperFlow helps you build a blog that keeps earning qualified demand, so growth is not tied to ad spend.
Citation-ready content
Customers don't visit sites (anymore). They ask ChatGPT, Claude, and Google AI. Your content needs to be citation-ready for the AI-search era.
Develop source-backed assets
Prompt-generated posts are easy to spot and easier to ignore. HarperFlow creates high quality articles that become your site's most valuable source-backed assets, stimulating long term growth.
Search is shifting to AI recommendations. HarperFlow turns topics into structured, authoritative content that earns top AI visibility automatically.
Real articles
## This is what we publish — and why AI cites it.
Every article on the HarperFlow blog is researched, written, cited, and published by our pipeline — no human edits. Here's a real one in full, next to the structure that gets AI answer engines to quote it.
✦
Built for AI citations
Structured so ChatGPT, Perplexity, and Google's AI can find, quote, and attribute it — not skim past it. Every item below is real in this article.
✓Answer-first opening (BLUF) + TL;DR
✓9 cited sources — verified URLs, search engines excluded
✓FAQ blocks, with the top four questions in FAQPage schema
✓Article JSON-LD — Article · Person · Organization
✓Internal links to related posts
✓Machine-readable dates + heading anchors
✓SEO meta + Open Graph + canonical URL
✓Listed in your site's llms.txt feed
Cited by AI · Bing Webmaster Tools
14,000+
documented AI citations of HarperFlow-published pages since we launched on August 1st—tracked in Bing Webmaster Tools.
[AI Visibility Score, Explained](https://www.harperflow.io/harperflow-blog/ai-visibility-score-explained-what-it-measures-and-how-to-raise-it)5.8k
[The Citation Techniques That Work](https://www.harperflow.io/harperflow-blog/ai-search-visibility-citation-techniques-that-actually-work)2.7k
[Structured Content Architecture](https://www.harperflow.io/harperflow-blog/structured-content-architecture-for-answer-engines-explained)1.3k
Source: Bing Webmaster Tools · updated September 27, 2026
2,889
Words
9
Cited sources
8
FAQ blocks
11
Min read
16
Content formats
Profiling
## We read your site, then write like you.
Before your first article, HarperFlow builds a brand profile from your live site — positioning, voice, audience, content pillars, even your accent color — and writes every post to match. Here's the profile we'd build for three very different brands.
Lumen Skincare
lumenskin.co
Ledgerwise
ledgerwise.io
Trailhead Supply
trailheadsupply.com
Illustrative examples — the kind of brand profile HarperFlow builds from your live site. Not real companies.
Positioning
Lumen is a clinical-skincare brand for people who want the science without the scare tactics.Every product is built around a single, well-studied active — retinaldehyde, niacinamide, azelaic acid — and the page explains exactly what it does, and what it doesn't.The positioning is deliberately anti-hype: no "10 years younger," no miracle claims, no fear-mongering about "toxins."Lumen competes on transparency instead — every formula lists its active concentration, and every claim points to a study.Its buyers are ingredient-literate; they've read the routines wikis and can tell a real INCI list from marketing filler.So the brand wins by respecting that literacy — the dermatologist friend who says retinol is worth it, but only if you introduce it slowly.It sits between drugstore basics and $150 luxury serums: premium, evidence-led, and honest about what a $38 bottle can and can't do.The competitive wedge is trust — Lumen would rather under-promise and keep a customer for a decade than over-claim and churn them in a month.Content sits at the top of that funnel: explain the science better than anyone, and the product becomes the obvious next step.And every article is written to be cited by the AI answer engines its research-first buyers now ask before they buy.
Voice & tone
WarmAuthoritativePlain-spokenFriendlyPlayfulTechnicalDirect
BeginnerPractitionerExpert
Brand accent
#E86A5E
Archetype · the trusted expert
Brand voice
Evidence-based, but never clinical-cold.Lumen writes like a knowledgeable friend with a dermatology background — warm, plain-spoken, unafraid to say "we don't know yet."Sentences stay short and declarative; jargon is defined the first time it appears.The brand never fear-mongers, never over-promises, and never talks down to the reader.Claims are hedged to what the evidence supports — "helps reduce," not "erases."Reassuring without being saccharine: closer to a trusted GP than a wellness influencer.
Audience
Skin-conscious millennials and older Gen-Z, roughly 25–40, who research ingredients before they buy.They've been burned by over-hyped products and are deeply skeptical of influencer marketing.They read ingredient lists, follow dermatologists, and want the "why" before spending $40 on a serum.Most are managing a specific concern — acne, texture, early aging, a compromised barrier.They want a methodical, evidence-led routine, not a 12-step ritual.They'll pay a premium for a brand that treats them as intelligent — and churn the moment one over-promises.Increasingly they open ChatGPT or Perplexity before Google — exactly where Lumen's content needs to surface.
Content types & formats16
Concept explainerHow-to guideTroubleshooting guideComparison tableDecision guideAnswer pageBuyer's guideDeep divePricing comparisonAlternatives roundupIntegration guideChecklistTemplate / resourceListicleUse caseCase study
Content pillars
Ingredient scienceBarrier healthRoutines by skin typeMyth-bustingActives & concentrationsSensitive-skin care
Image style6
PhotographicFlat IllustrationSoft 3DLine ArtCollageIsometric
Topics it should own12
Retinol vs retinaldehydeRepairing a damaged barrierNiacinamide, actually explainedThe right order to layer activesSPF for oily, acne-prone skinAzelaic acid for rednessDo you need a vitamin C serum?Introducing retinoids without peelingFungal acne vs regular acneReading an INCI listHyaluronic acid hydration mythsBuilding a minimalist routine
Article template6
StoryMagazine split heroGuideCentred field-guideSidebarSticky rail + authorFull-bleedEdge-to-edge heroConversionTakeaways + CTA panelCinematicFull-scrim title
Author card
Dr. Rae Okafor
Consultant Dermatologist · Lumen
Board-certified dermatologist who writes and reviews Lumen's evidence explainers.
inxweb
name · role @ company · blurb · avatar · up to 2 socials
CTA / callout card
Not sure where to start?
Take the 60-second skin quiz and get a routine built around your actual concern — not a 12-step ritual.
Start the skin quiz
header · body · button text · link
Positioning
Ledgerwise is close-automation software for finance teams tired of month-end taking two weeks.It connects to the ERP, the bank feeds, and the sub-ledgers, then reconciles and flags only the exceptions a human actually needs to see.The positioning is unapologetically for practitioners: it assumes you know what an accrual is and never dumbs the product down.It competes on trust and auditability — every automated entry carries a full trail, so controllers can defend it without a spreadsheet archaeology dig.The wedge against legacy suites is speed-to-value: live in a week, not a six-month implementation with a consulting firm attached.Buyers are controllers at 50–500-person companies who have outgrown QuickBooks but refuse to be held hostage by a NetSuite rollout.They're skeptical of "AI bookkeeping" hype and want to see exactly what the software touched, and why.So the content strategy is proof, not promise: show the close checklist, the reconciliation logic, the integration mechanics.Ledgerwise earns the demo by being the most useful explainer of the modern close on the internet.And it writes for the AI engines finance buyers now ask to shortlist tools before they ever book a call.
Voice & tone
WarmAuthoritativePlain-spokenFriendlyPlayfulTechnicalDirect
BeginnerPractitionerExpert
Brand accent
#4B5AE4
Archetype · the seasoned operator
Brand voice
Precise, calm, and quietly confident — the senior controller who has closed a thousand books.Technical without being cold; it respects the reader's expertise and never over-explains the basics.Direct and specific: real numbers, real workflows, no vague "streamline your finances" filler.Honest about limits — what to automate, and what a human should still review.Authoritative but never salesy; the product is the proof, so the writing doesn't oversell.Reads at an expert level — this is B2B, and the audience does this for a living.
Audience
Controllers, accounting managers, and finance leads at 50–500-employee companies.They own the monthly close and are judged on how fast and how clean it is.They've outgrown entry-level tools but dread a multi-quarter ERP implementation.Deeply skeptical of "AI does your books" marketing; they want auditable, explainable automation.They evaluate tools like engineers: integrations, controls, edge cases, and the audit trail.They'll champion a tool only if it survives their CFO's and auditor's scrutiny.Increasingly they ask ChatGPT and Perplexity to shortlist close-automation tools — where Ledgerwise needs to be the cited answer.
Content types & formats16
Concept explainerHow-to guideTroubleshooting guideComparison tableDecision guideAnswer pageBuyer's guideDeep divePricing comparisonAlternatives roundupIntegration guideChecklistTemplate / resourceListicleUse caseCase study
Content pillars
The monthly closeReconciliation & controlsERP & bank integrationsAudit & complianceFinancial reportingSpend management
Image style6
PhotographicFlat IllustrationSoft 3DLine ArtCollageIsometric
Topics it should own12
How to close the books in 5 daysAutomating bank reconciliationNetSuite vs Ledgerwise for mid-marketAn audit-ready close checklistAccruals vs cash: what to automateConnecting your ERP without breaking itMonth-end bottlenecks to cut firstQuickBooks to Ledgerwise migrationSOX controls for a lean teamFlux analysis, explainedChoosing close-automation softwareReducing manual journal entries
Article template6
StoryMagazine split heroGuideCentred field-guideSidebarSticky rail + authorFull-bleedEdge-to-edge heroConversionTakeaways + CTA panelCinematicFull-scrim title
Author card
Marcus Chen
Controller-in-Residence · Ledgerwise
Former Big-Four auditor; writes Ledgerwise's close and controls playbooks.
inxweb
name · role @ company · blurb · avatar · up to 3 socials
CTA / callout card
See a 5-day close
Book a 20-minute demo and watch Ledgerwise reconcile a real month in minutes.
Book a demo
header · body · button text · link
Positioning
Trailhead Supply is a direct-to-consumer outdoor brand for people who actually sleep in the dirt, not just shop the aesthetic.It sells backpacking gear chosen for the trail, not the trend — ultralight, field-tested, and honestly rated.The positioning is anti-gimmick: no "extreme" marketing, no gear you'll use once, just kit that earns its weight in your pack.It competes on credibility — every product page tells you what the gear is bad at, not just what it's good at.Buyers are experienced hikers who spot marketing fluff instantly and want specifics: denier, fill power, packed weight, real durability.So the brand wins by being the most trustworthy gear advisor — the friend who's already hiked the trail you're planning.Content is the trailhead of that relationship: help people choose right, and the gear sale follows naturally.It sits between big-box outdoor retailers and boutique cottage brands: curated, expert, and refreshingly plain-spoken.Trust compounds — a hiker who buys the right tent comes back for the pack, the quilt, and the stove.And every guide is written to be the answer an AI engine gives when someone asks what to actually carry.
Voice & tone
WarmAuthoritativePlain-spokenFriendlyPlayfulTechnicalDirect
BeginnerPractitionerExpert
Brand accent
#2F7D4F
Archetype · the trusted guide
Brand voice
Warm, plain-spoken, and a little wry — like advice from a friend at the trailhead.Knowledgeable but never gatekeeping; it welcomes beginners without dumbing things down.Specific and honest: real weights, real trade-offs, and the downsides stated plainly.Encouraging without hype — it gets you outside, it doesn't sell you a fantasy.Playful where it fits, serious where safety matters.Reads like a trusted guide, not a catalog.
Audience
Backpackers, thru-hikers, and weekend adventurers, roughly 25–45, who research gear obsessively.They've been burned by heavy, overpriced kit and want honest, weight-conscious recommendations.They read spec sheets, compare packed weights, and trust field reviews over ad copy.Many are planning a specific trip — a first thru-hike, a winter overnight, an alpine start.They value durability and repairability and will pay for gear that lasts a decade.They're loyal to brands that tell the truth and quick to call out ones that don't.They ask communities before buying — and increasingly ask AI engines what to pack, where Trailhead's guides need to get cited.
Content types & formats16
Concept explainerHow-to guideTroubleshooting guideComparison tableDecision guideAnswer pageBuyer's guideDeep divePricing comparisonAlternatives roundupIntegration guideChecklistTemplate / resourceListicleUse caseCase study
Content pillars
Gear guidesTrail skillsTrip planningGear maintenanceUltralight principlesBeginner basics
Image style6
PhotographicFlat IllustrationSoft 3DLine ArtCollageIsometric
Topics it should own12
How to choose a backpacking tentUltralight vs comfort trade-offsDown vs synthetic sleeping bagsBuilding your first thru-hike kitSizing a pack that actually fitsWater filtration on the trail, comparedWhat to pack for a winter overnightTrail-runner vs boot for distanceRepairing gear in the fieldCutting pack weight safelyLayering systems, explainedBackcountry first-aid essentials
Article template6
StoryMagazine split heroGuideCentred field-guideSidebarSticky rail + authorFull-bleedEdge-to-edge heroConversionTakeaways + CTA panelCinematicFull-scrim title
Author card
Sierra Watts
Head Guide · Trailhead Supply
AT and PCT thru-hiker; writes Trailhead's field-tested gear guides.
inxweb
name · role @ company · blurb · avatar · up to 3 socials
CTA / callout card
Not sure what to pack?
Take the 2-minute trip quiz and get a gear list built for your route and season.
Build my kit
header · body · button text · link
### Built for depth, accuracy, and seamless publishing.
HarperFlow replaces the endless ad spend with a GEO publishing system: it finds the opportunities, crafts the answers, cites the sources, and learns from every article, so your blog compounds while you run the business.
## Our HarperFlow Blog
Every article on our site was researched, cited, and published using HarperFlow. See what full-pipeline automation looks like in production.
[ Structured Content Architecture for Answer Engines September 26, 2026 · 8 min read
This article defines an accordion design template as a stack of five parts—container, header trigger, label, state icon, and panel—and explains single- versus multi-expand behavior. It maps patterns to jobs like FAQ, pricing, specs, and navigation, details W3C ARIA requirements for buttons, aria-expanded, and keyboard interaction, shows a no-code Webflow build, and advises when to avoid accordions for scannability and SEO. Read article →](https://www.harperflow.io/harperflow-blog/accordion-design-template-anatomy-patterns-setup)
## Latest articles
[ Structured Content Architecture for Answer Engines
Semantic Content Architecture: A Practical Framework
This article defines semantic content architecture as organizing content around entities, typed relationships, and machine-readable markup. It distinguishes it from topic clusters, information architecture, and schema markup, details four relationship types—hierarchical, sequential, contrastive, dependency—and lays out a build sequence from entity inventory to bidirectional linking and JSON-LD. It closes with common legibility failures and maintenance habits for scaling. September 26, 2026 · 10 min Read more →](https://www.harperflow.io/harperflow-blog/semantic-content-architecture-practical-framework)
[ CMS Integration and Management (Webflow, WordPress, Shopify)
How to Build an Accordion in Webflow (3 Methods Compared)
This guide compares three no-code ways to build an accordion in Webflow: native Dropdown repurposed, HTML details/summary, and Div blocks with Interactions. It evaluates setup effort, accessibility, animation smoothness, and design flexibility, then provides step-by-step build instructions, ARIA and keyboard requirements, CMS binding tips, SEO safety, and fixes for common overlap and height bugs. Dropdown wins for speed; other methods suit semantics or custom design. September 26, 2026 · 12 min Read more →](https://www.harperflow.io/harperflow-blog/how-to-build-an-accordion-in-webflow-3-methods-compared)
[ Structured Content Architecture for Answer Engines
What Is LLM-Friendly Content? A Practical Definition
This article defines llm friendly content as structure that lets AI systems parse, chunk, and cite discrete passages, not just rank. It explains crawler access via GPTBot and Google-Extended, the llms.txt convention, schema.org and semantic HTML signals, how RAG chunking selects answers, shows a before-and-after rewrite, and ends with a pass/fail checklist to verify every page remains citable. September 26, 2026 · 8 min Read more →](https://www.harperflow.io/harperflow-blog/what-is-llm-friendly-content-a-practical-definition)
## A complete publishing engine.
HarperFlow takes articles from research and citations through scheduling and publication on your Webflow, WordPress, or Shopify blog.
HarperFlow Prompt Writers
HarperFlow turns each article into a structured GEO asset: researched, cited, formatted, mapped, published, and connected to the rest of your content engine.
Research, citations, and GEO structure included
Direct publishing to Webflow, WordPress, and Shopify
Topic memory, internal links, and weekly flow
Every article is built to become part of a larger AI-search growth engine, not another isolated blog post.
Generic AI writers can produce words quickly, but they rarely solve the harder parts: depth, source discipline, structure, and automated publishing.
Every article starts from scratch
You still need to verify all the facts
Publishing remains a manual weekly chore
The output may look complete, but your team still has to make it publishable, credible, and useful.
## See how HarperFlow compares
Honest, sourced side-by-sides against the tools you're weighing — we even name where they beat us.
HarperFlow vs Soro
The cheaper, more established auto-publisher — HarperFlow adds citation rigour, an independent audit, and a veto on every post.
AI PUBLISHER
[Read the comparison →](https://www.harperflow.io/compare/soro)
![Abstract logo with two shapes: orange rounded tops and black triangular bases.]()
HarperFlow vs BabyLoveGrowth
The feature-richest rival — backlinks and free tools included. HarperFlow wins on citation, research quality, and control.
AI PUBLISHER
[Read the comparison →](https://www.harperflow.io/compare/babylovegrowth)
HarperFlow vs Surfer SEO
Surfer grades drafts against the SERP; HarperFlow writes, cites, audits, and publishes finished articles built to be cited by AI.
SEO TOOL
[Read the comparison →](https://www.harperflow.io/compare/surferseo)
HarperFlow vs Jasper
Jasper is a versatile AI writing tool. HarperFlow is a service that researches, audits, and publishes your blog articles for you.
AI WRITER
[Read the comparison →](https://www.harperflow.io/compare/jasper)
[See all 10 comparisons
](https://www.harperflow.io/compare)
Languages & markets
## Written in your market's language — and tuned to what it searches for.
HarperFlow researches and writes natively in English, Spanish, German, and French, across roughly twenty search markets. Just as important: every keyword and topic is scored against live search-volume data for your business's main country or target market — so you win demand that actually exists there, not generic global terms.
4
Native-grade languages
\~20
Search markets
English
English
Search markets
US
UK
CA
AU
IE
NZ
Español
Spanish
Search markets
ES
MX
AR
CO
CL
US
Deutsch
German
Search markets
DE
AT
CH
Français
French
Search markets
FR
CA
BE
CH
Tuned to your target market, not the global average
We pull live search-volume and competition data per country, so an English site selling into the UK gets UK-weighted keywords and topics — never US ones by default. The same market-tuning runs for every language and country above.
More languages and markets are coming soon.
Free trial available
## Webflow & WordPress pricing
Choose 5–30 articles a month for Webflow or WordPress. Every volume includes the full publishing pipeline. Shopify uses separate pricing from $39.90/month, billed through Shopify.
[Shopify pricing and installation →](https://www.harperflow.io/shopify)
Webflow & WordPress — you set the volume
−
10
articles / month
\+
5 · $62/mo
30 · $297/mo
5102030
$119/month
$11.90 per article
Published about every 3 days, hands-free.
[Start with 10 articles / month](https://app.harperflow.io/auth/register?credits=10)
Renews at your amount until you change it. Upgrade, downgrade, or cancel anytime — extra credits available mid-month at your rate.
What you're paying for
### Every article ships with the full engine.
No feature tiers, no gates — a 5-article package gets the same machinery as a 30-article one.
✓
8–12 cited sources per article
✓
Source-to-snippet hooks
✓
16 article formats
✓
Internal link creation
✓
Author SEO cards & styled callouts
✓
FAQ blocks
✓
JSON-LD schema
✓
Written in your brand voice
✓
Veto anything before it's live
✓
Run your blog on autopilot
Connect your site once — HarperFlow discovers, writes, cites, links, and publishes on schedule from then on.
Agencies & Enterprise
### Running content for many sites? Plans start at $1,020/month.
✓
Custom article volume across all your client sites
✓
One consolidated monthly bill for every site
✓
Volume rates as low as $6.00 per article
✓
Dedicated pipeline — bring your preferred AI model
[Book a call with the founder](https://cal.com/hesham.us/harperflow-agencies-enterprise)
Arranged bespoke, 1-1 with HarperFlow's founder — 120 articles included at the entry tier.
HarperFlow pricing: a monthly subscription of 5 to 30 GEO-optimized, autopilot-published articles for one website. The price per article decreases as the package grows: $12.90 minus $0.10 for each article in the package. Agency and Enterprise plans for multiple websites start at $1,020 per month (120 articles included, volume rates as low as $6.00 per article) and are arranged through a call with the founder.
| Articles per month | Price per article | Monthly price (USD) |
| --- | --- | --- |
| 5 articles | $12.40 per article | $62 per month |
| 6 articles | $12.30 per article | $73.80 per month |
| 7 articles | $12.20 per article | $85.40 per month |
| 8 articles | $12.10 per article | $96.80 per month |
| 9 articles | $12 per article | $108 per month |
| 10 articles | $11.90 per article | $119 per month |
| 11 articles | $11.80 per article | $129.80 per month |
| 12 articles | $11.70 per article | $140.40 per month |
| 13 articles | $11.60 per article | $150.80 per month |
| 14 articles | $11.50 per article | $161 per month |
| 15 articles | $11.40 per article | $171 per month |
| 16 articles | $11.30 per article | $180.80 per month |
| 17 articles | $11.20 per article | $190.40 per month |
| 18 articles | $11.10 per article | $199.80 per month |
| 19 articles | $11 per article | $209 per month |
| 20 articles | $10.90 per article | $218 per month |
| 21 articles | $10.80 per article | $226.80 per month |
| 22 articles | $10.70 per article | $235.40 per month |
| 23 articles | $10.60 per article | $243.80 per month |
| 24 articles | $10.50 per article | $252 per month |
| 25 articles | $10.40 per article | $260 per month |
| 26 articles | $10.30 per article | $267.80 per month |
| 27 articles | $10.20 per article | $275.40 per month |
| 28 articles | $10.10 per article | $282.80 per month |
| 29 articles | $10 per article | $290 per month |
| 30 articles | $9.90 per article | $297 per month |
## Earn AI-search demand with 18+ features
We built HarperFlow for our clients—then stacked it with features.
[Explore the docs Setup, autopilot, billing, and GEO guides, Browse the docs](https://www.harperflow.io/docs)
Internal Link Creation
Finds natural internal linking moments so new articles strengthen the wider content library over time.
Rich Article Formatting
Publishes structured articles with headings, tables, FAQs, citations, and images to Webflow, WordPress, and Shopify blogs.
Sixteen Article Formats
Matches each topic to the right format, from answer pages to comparisons, checklists, and deep dives.
SEO Metadata Generation
Generates article titles, descriptions, slugs, and metadata for your publishing platform.
TL;DR Takeaways
Adds punchy article takeaways that make long-form content easier to scan, remember, and cite quickly.
Source-to-Snippet Hooks
Builds concise, source-backed passages designed to be lifted, quoted, and reused by AI answer engines.
Content Calendar Automation
Turns approved opportunities into a publishing rhythm without another spreadsheet or editorial chase cycle.
Styled Content Callouts
Highlights key ideas, warnings, summaries, and proof points with polished blocks inside the article body.
Featured Image Generation
Creates relevant article images and alt text for your connected blog.
Brand Voice Memory
Keeps articles aligned with your positioning, audience, tone, product rules, and content pillars every time.
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/features
# Everything your blog needs to earn AI-search demand
Built first for Webflow, HarperFlow now researches, writes, cites, schedules, and publishes articles to Webflow, WordPress, and Shopify blogs. One publishing engine, with integrations adapted to each platform.
## A complete publishing engine.
HarperFlow takes articles from research and citations through scheduling and publication on your Webflow, WordPress, or Shopify blog.
HarperFlow Prompt Writers
HarperFlow turns each article into a structured GEO asset: researched, cited, formatted, mapped, published, and connected to the rest of your content engine.
Research, citations, and GEO structure included
Direct publishing to Webflow, WordPress, and Shopify
Topic memory, internal links, and weekly flow
Every article is built to become part of a larger AI-search growth engine, not another isolated blog post.
Generic AI writers can produce words quickly, but they rarely solve the harder parts: depth, source discipline, structure, and automated publishing.
Every article starts from scratch
You still need to verify all the facts
Publishing remains a manual weekly chore
The output may look complete, but your team still has to make it publishable, credible, and useful.
## Outsmart the changing search landscape.
Search is shifting to AI recommendations. HarperFlow turns topics into structured, authoritative content that earns top AI visibility automatically.
[Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Every paid click resets the meter. HarperFlow helps you build a blog that keeps earning qualified demand, so growth is not tied to ad spend.
Customers don't visit sites (anymore). They ask ChatGPT, Claude, and Google AI. Your content needs to be answer-ready and citation-ready for the AI-search era.
Prompt-generated posts are easy to spot and easier to ignore. HarperFlow creates high quality articles that become your site's most valuable source-backed assets.
## Earn AI-search demand with 18+ features
We built HarperFlow for our clients—then stacked it with more features.
Opportunity Discovery
Finds high-intent topics your customers are already asking search engines, AI assistants, and competitors.
AI-Search Gap Scanning
Spots answer gaps where your site can become the source AI engines quote before competitors.
Topic Cluster Planning
Groups articles around connected themes, so your blog builds authority on specific topic clusters.
Content Calendar Automation
Turns approved opportunities into a publishing rhythm without another spreadsheet or editorial chase cycle.
Sixteen Article Formats
Matches each topic to the right format, from answer pages to comparisons, checklists, and deep dives.
Brand Voice Memory
Keeps articles aligned with your positioning, audience, tone, product rules, and content pillars every time.
Source-to-Snippet Hooks
Builds concise, source-backed passages designed to be lifted, quoted, and reused by AI answer engines.
Citation Architecture
Connects claims, sources, FAQs, and summaries into a cleaner trail for readers and AI engines.
TL;DR Takeaways
Adds punchy article takeaways that make long-form content easier to scan, remember, and cite quickly.
FAQ Blocks For Every Article
Includes structured FAQs that capture long-tail questions and give AI engines cleaner answer targets inside articles.
Styled Content Callouts
Highlights key ideas, warnings, summaries, and proof points with polished blocks inside the article body.
Rich Article Formatting
Publishes headings, lists, tables, FAQs, citations, and images to your Webflow, WordPress, or Shopify blog.
Featured Image Generation
Creates article images and alt text for your connected blog, with presentation adapted to your platform.
SEO Metadata Generation
Generates titles, descriptions, slugs, and article metadata for your publishing platform.
Internal Link Creation
Finds natural internal linking moments so new articles strengthen the wider content library over time.
SEO-Optimised Author Cards
Adds bylines and author context that support credibility, presentation, and on-page article completeness.
Article Themes Included
Applies polished article themes so automated posts feel native, styled, and ready to publish instantly.
33-Field CMS Structure
For Webflow, maps articles into a purpose-built 33-field CMS collection. WordPress and Shopify use their own publishing integrations.
show all features
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/pricing
# Pricing Plans
Choose 5–30 articles a month with the full publishing pipeline. Webflow and WordPress use web billing; Shopify uses its own pricing and Shopify Billing.
Free trial available
## Webflow & WordPress pricing
Choose 5–30 articles a month for Webflow or WordPress. Every volume includes the full publishing pipeline. Shopify uses separate pricing from $39.90/month, billed through Shopify.
[Shopify pricing and installation →](https://www.harperflow.io/shopify)
Webflow & WordPress — you set the volume
−
10
articles / month
\+
5 · $62/mo
30 · $297/mo
5102030
$119/month
$11.90 per article
Published about every 3 days, hands-free.
[Start with 10 articles / month](https://app.harperflow.io/auth/register?credits=10)
Renews at your amount until you change it. Upgrade, downgrade, or cancel anytime — extra credits available mid-month at your rate.
What you're paying for
### Every article ships with the full engine.
No feature tiers, no gates — a 5-article package gets the same machinery as a 30-article one.
✓
8–12 cited sources per article
✓
Source-to-snippet hooks
✓
16 article formats
✓
Internal link creation
✓
Author SEO cards & styled callouts
✓
FAQ blocks
✓
JSON-LD schema
✓
Written in your brand voice
✓
Veto anything before it's live
✓
Run your blog on autopilot
Connect your site once — HarperFlow discovers, writes, cites, links, and publishes on schedule from then on.
Agencies & Enterprise
### Running content for many sites? Plans start at $1,020/month.
✓
Custom article volume across all your client sites
✓
One consolidated monthly bill for every site
✓
Volume rates as low as $6.00 per article
✓
Dedicated pipeline — bring your preferred AI model
[Book a call with the founder](https://cal.com/hesham.us/harperflow-agencies-enterprise)
Arranged bespoke, 1-1 with HarperFlow's founder — 120 articles included at the entry tier.
HarperFlow pricing: a monthly subscription of 5 to 30 GEO-optimized, autopilot-published articles for one website. The price per article decreases as the package grows: $12.90 minus $0.10 for each article in the package. Agency and Enterprise plans for multiple websites start at $1,020 per month (120 articles included, volume rates as low as $6.00 per article) and are arranged through a call with the founder.
| Articles per month | Price per article | Monthly price (USD) |
| --- | --- | --- |
| 5 articles | $12.40 per article | $62 per month |
| 6 articles | $12.30 per article | $73.80 per month |
| 7 articles | $12.20 per article | $85.40 per month |
| 8 articles | $12.10 per article | $96.80 per month |
| 9 articles | $12 per article | $108 per month |
| 10 articles | $11.90 per article | $119 per month |
| 11 articles | $11.80 per article | $129.80 per month |
| 12 articles | $11.70 per article | $140.40 per month |
| 13 articles | $11.60 per article | $150.80 per month |
| 14 articles | $11.50 per article | $161 per month |
| 15 articles | $11.40 per article | $171 per month |
| 16 articles | $11.30 per article | $180.80 per month |
| 17 articles | $11.20 per article | $190.40 per month |
| 18 articles | $11.10 per article | $199.80 per month |
| 19 articles | $11 per article | $209 per month |
| 20 articles | $10.90 per article | $218 per month |
| 21 articles | $10.80 per article | $226.80 per month |
| 22 articles | $10.70 per article | $235.40 per month |
| 23 articles | $10.60 per article | $243.80 per month |
| 24 articles | $10.50 per article | $252 per month |
| 25 articles | $10.40 per article | $260 per month |
| 26 articles | $10.30 per article | $267.80 per month |
| 27 articles | $10.20 per article | $275.40 per month |
| 28 articles | $10.10 per article | $282.80 per month |
| 29 articles | $10 per article | $290 per month |
| 30 articles | $9.90 per article | $297 per month |
## Earn AI-search demand with 18+ features
We built HarperFlow for our clients—then stacked it with more features.
Opportunity Discovery
Finds high-intent topics your customers are already asking search engines, AI assistants, and competitors.
AI-Search Gap Scanning
Spots answer gaps where your site can become the source AI engines quote before competitors.
Topic Cluster Planning
Groups articles around connected themes, so your blog builds authority on specific topic clusters.
Content Calendar Automation
Turns approved opportunities into a publishing rhythm without another spreadsheet or editorial chase cycle.
Sixteen Article Formats
Matches each topic to the right format, from answer pages to comparisons, checklists, and deep dives.
Brand Voice Memory
Keeps articles aligned with your positioning, audience, tone, product rules, and content pillars every time.
Source-to-Snippet Hooks
Builds concise, source-backed passages designed to be lifted, quoted, and reused by AI answer engines.
Citation Architecture
Connects claims, sources, FAQs, and summaries into a cleaner trail for readers and AI engines.
TL;DR Takeaways
Adds punchy article takeaways that make long-form content easier to scan, remember, and cite quickly.
FAQ Blocks For Every Article
Includes structured FAQs that capture long-tail questions and give AI engines cleaner answer targets inside articles.
Styled Content Callouts
Highlights key ideas, warnings, summaries, and proof points with polished blocks inside the article body.
Rich Article Formatting
Publishes headings, lists, tables, FAQs, citations, and images to your Webflow, WordPress, or Shopify blog.
Featured Image Generation
Creates article images and alt text for your connected blog, with presentation adapted to your platform.
SEO Metadata Generation
Generates titles, descriptions, slugs, and article metadata for your publishing platform.
Internal Link Creation
Finds natural internal linking moments so new articles strengthen the wider content library over time.
SEO-Optimised Author Cards
Adds bylines and author context that support credibility, presentation, and on-page article completeness.
Article Themes Included
Applies polished article themes so automated posts feel native, styled, and ready to publish instantly.
33-Field CMS Structure
For Webflow, maps articles into a purpose-built 33-field CMS collection. WordPress and Shopify use their own publishing integrations.
show all features
## Frequently Asked Questions
Here are the answers to the things most businesses ask before using HarperFlow to publish researched, source-backed articles.
What does HarperFlow actually do?
HarperFlow researches, writes, formats, and publishes source-backed articles to Webflow, WordPress, and Shopify blogs. It handles research, citations, article structure, scheduling, and publishing, with a veto before scheduled articles go live. Articles are built to be cited by AI; citations, rankings, and traffic are not guaranteed.
Is HarperFlow just another AI writing tool?
No. Prompt writers create drafts. HarperFlow runs a publishing pipeline. It handles opportunity discovery, research, citations, formatting, metadata, CMS mapping, internal links, content calendars, and brand memory so your blog can compound over time.
How do I know the articles are high quality?
HarperFlow was built by a content-obsessed automation lab, and we use the same pipeline on our own blog and sister brands. Each article is researched, structured, source-backed, and built with the standards we would use if writing it ourselves.
What platforms does HarperFlow support?
HarperFlow supports Webflow, WordPress, and Shopify. Install through the Webflow App Marketplace, our WordPress.org plugin, or the Shopify App Store. Shopify is billed through Shopify from $39.90/month with a 7-day free trial. Webflow and WordPress start at $62/month. Wix is in development.
Will HarperFlow publish without me knowing?
No. Every Monday, you receive a simple digest of the articles scheduled for that week, so you can see what is coming without managing the workflow yourself.
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/shopify
# Your Shopify blog on autopilot.
Your products have a story. Give your blog the time to tell it. HarperFlow researches, writes, and publishes source-backed articles in your brand voice—straight to your Shopify blog.
[Install from the Shopify App Store](https://apps.shopify.com/harperflow) [Read the setup guide](https://www.harperflow.io/docs/connect-shopify)
From $39.90/month · 7-day free trial · Billed through Shopify
YOUR BRAND. YOUR BLOG.
## A complete article, already on your calendar.
Built to be cited by AI. Published in your Shopify theme.
- Source-backed answers
- Images, tables, and FAQs
- Your voice and internal links
- A schedule you can veto
Research
→
Write
→
Publish
ONE LESS THING TO RUN
## From your store to your next article.
01
### Connect your store
Install HarperFlow, connect your account, and choose an existing blog or a dedicated HarperFlow blog.
02
### Let it learn your business
HarperFlow learns your store, audience, and brand voice, then plans topics around the questions your customers ask.
03
### Keep the final say
Articles receive a publishing slot and an email digest first. Skip anything you do not want published; the rest goes live on schedule.
## The whole publishing job, handled.
### Written for your store
Your positioning, products, audience, and voice shape the article. You can refine the profile HarperFlow uses.
### Structured for readers and AI
Cited sources, clear answers, headings, images, tables, FAQs, and metadata give each article a complete structure.
### Published in your own blog
Content goes into your selected Shopify blog and uses your theme’s article layout. No weekly copy-and-paste routine.
[Explore articles made with HarperFlow](https://www.harperflow.io/harperflow-blog)
Built to be cited is a design goal, not a guarantee of citations, rankings, or traffic.
SHOPIFY PRICING
## Choose the pace. Keep the full engine.
Choose 5–30 articles per month. Research, writing, structure, images, and scheduled publishing are included at every volume.
5 articles / month
$39.90
per month · USD
10 articles / month
$74.90
per month · USD
20 articles / month
$134.90
per month · USD
30 articles / month
$184.90
per month · USD
[Start your 7-day trial](https://apps.shopify.com/harperflow)
Shopify Billing. Review the trial terms and recurring charge in Shopify before approving. Webflow and WordPress use separate pricing.
[Compare platform pricing](https://www.harperflow.io/pricing)
## A few things worth knowing.
### Does HarperFlow publish directly to Shopify?
Yes. After installation and setup, HarperFlow publishes scheduled articles to the Shopify blog you select. Articles use your store’s existing theme.
### Do I need to approve every article?
No. You receive a veto digest before scheduled publication and can skip an article. Articles you do not veto publish automatically at their scheduled time.
### How do billing and cancellation work?
Shopify handles subscription and add-on charge approvals. Change your article volume in HarperFlow’s Plan page and approve it in Shopify. Uninstalling the app stops future billing cycles; charges already incurred may still be payable.
### Does it guarantee AI citations?
No. HarperFlow creates source-backed, structured articles built to be cited by AI. Search engines and AI systems decide what to rank, retrieve, and cite.
[More installation and billing help](https://www.harperflow.io/docs/connect-shopify)
## Put your Shopify blog on autopilot.
Make room for the work only you can do.
[Install HarperFlow for Shopify](https://apps.shopify.com/harperflow)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/about
# Built for the businesses AI search should find
[Start Your Free Trial Today Get started now](https://www.harperflow.io/contact)
HarperFlow was built by Hesham Mashhour — a University of Cambridge graduate and doctor turned creator, turned content-automation engineer.
## Our story
HarperFlow started from a simple frustration: most businesses either overpay for attention or publish content that adds nothing.
AI search made that problem sharper. If customers ask
answer engines before they visit websites, your blog cannot be a dumping ground for generic posts. It has to become the source those engines can understand, quote, and recommend.
HarperFlow began with Webflow. Today, the same publishing engine researches topics, writes source-backed articles, adds structure and images, schedules publication, and publishes to Webflow, WordPress, and Shopify blogs. Each integration handles delivery in the way its platform expects.
We use the same pipeline on our own blog and sister brands before offering it to customers.
117
AI answers analyzed across ChatGPT, Perplexity, and Gemini in our first study
98%
of those answers named a specific brand — the visibility we build for
The founder
## Built by someone who publishes for a living.
Hesham Mashhour
Founder, HarperFlow · University of Cambridge graduate · doctor turned creator
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Book a call](https://cal.com/hesham.us/)
Hesham Mashhour is a University of Cambridge graduate and medical doctor who left medicine to build on the internet. His YouTube channel, NeuroEverything, grew past 40,000 subscribers — and making content at that scale is what turned him into an automation engineer.
He kept building the tools he wished existed, until one of them became HarperFlow — the publishing pipeline he now runs on his own sites before offering it to anyone who would rather build than market themselves.
Also built by Hesham
[
NeuroEverything ↗
A YouTube channel on the brain and how to get more from yours — past 40,000 subscribers.](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [
Nodey ↗
Open-source n8n community nodes that wire automation into tools like vidIQ and Ghost.](https://getnodey.com) [
NanoWhite ↗
A consumer brand run on the same publish-it-yourself playbook.](https://nanowhite.us) [
Skinkins ↗
Another direct-to-consumer brand, and a live test of the system.](https://skinkins.com) [
Books ↗
Published books — the long-form habit behind the whole pipeline.](https://www.amazon.com/Hesham-Mashhour/e/B0BJ7SW5VR) [
Better Brain Lab ↗
The studio all of it ships under.](https://betterbrainlab.com)
## Our values
Research Before Output
Every article starts with opportunities, sources, and evidence before a single sentence is drafted properly.
Structure Creates Visibility
Great ideas need clean architecture, metadata, headings, FAQs, and formats AI engines can understand quickly.
Sources Over Sludge
Claims should be traceable, citation-ready, and useful enough for readers and answer engines to trust.
Operators Deserve Ease
Business owners should get compounding content without managing prompts, editors, calendars, or CMS chores weekly.
Compound, Don’t Chase
Content should keep building authority, internal links, and demand long after publish day passes quietly.
Raising Standards
Automation is only worth using when it improves quality, consistency, speed, and operational focus together.
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/geo-website-audit
# Free GEO Audit & AI Visibility Checker
Run a free GEO audit — an AI visibility checker for your site — and see where your content can become easier for AI engines like ChatGPT and Perplexity to understand, quote, and recommend.
What we check
## What the audit actually looks at
Six things AI engines weigh when they decide whether your page is worth quoting — the same dimensions our own research measures.
01
### Structure and headings
Whether your pages use the clear question-and-answer shape AI engines lift from. In our study, structured pages with FAQ blocks and tables were cited 75% of the time versus 22% without — the single strongest predictor we measured.
02
### Citations and sourcing
Whether your claims trace to named sources an engine can verify. Content that grounds its numbers is content an assistant can repeat with confidence.
03
### Structured data
Whether your pages ship valid Article, FAQ, and Organization schema — the metadata layer engines read before they read your prose.
04
### Freshness and dating
Whether your content is visibly dated and maintained. Engines prefer sources they can time-stamp; undated pages read as stale.
05
### Snippet-worthiness
Whether an assistant could lift a clean, self-contained answer from your page — short verdicts, definitions, and takeaways an answer engine can quote verbatim.
06
### AI crawler access
Whether AI search bots can actually reach and read your pages — robots.txt configuration, rendering, and the basics that silently exclude sites from AI answers.
How it works
## Paste a URL. Get a grade.
No signup to see your result — the full report lands in your inbox.
01
### Paste your website above
One URL is all the audit needs. It runs in seconds against your live site — not a cached copy.
02
### We read your site the way AI does
The audit evaluates your pages the way ChatGPT, Perplexity, and Google AI read them — structure, sources, schema, and access — and turns that into an AI-visibility grade.
03
### Get your grade and a fix list
You see your grade instantly, and the full report — what's holding your site back and the prioritized fixes — is delivered to your inbox.
Why it matters
## Your buyers are already asking AI
Everything below is measured — from our own published, recomputable research.
### AI answers name brands, not links
When we put 39 real buyer questions through ChatGPT, Perplexity, and Gemini, [98% of the answers named a specific brand](https://www.harperflow.io/state-of-ai-citations). Buyers get a short list of names — and if AI isn't naming you, you're not losing a ranking, you're invisible.
### Structure decides who gets cited
The strongest predictor of whether a brand's own site gets cited isn't authority or age — it's [how the page is built](https://www.harperflow.io/can-your-site-get-cited-by-ai). Structured pages were cited 75% of the time versus 22% without, a statistically significant gap (p=0.004). That's exactly what this audit grades.
### The landscape shifts fast
Fifteen days after our first measurement, Reddit collapsed from 87% of Perplexity's citations to [18%](https://www.harperflow.io/the-reddit-collapse). AI visibility isn't a one-time project — it's a surface you maintain. The audit tells you where you stand today.
Questions
## The free GEO audit, answered
Is the audit really free?
Yes — completely. You see your AI-visibility grade instantly, with no signup required to view it, and the full report is emailed to you at no cost.
What exactly does the audit check?
Six dimensions of AI readability: page structure and headings, citation and sourcing habits, structured data (schema), content freshness, snippet-worthiness, and whether AI crawlers can access your pages at all.
How long does it take?
Seconds. The audit runs against your live site the moment you submit your URL, and your grade appears right away. The detailed report follows to your inbox.
Do I need a HarperFlow account?
No. The audit is a standalone tool. If the results convince you the gap is worth closing, HarperFlow's publishing pipeline is the fix — but the audit stands on its own.
What should I do with the results?
Start with the highest-impact items — usually structure and schema. If you'd rather the fixes publish themselves, that's what HarperFlow's autopilot was built for: cited, structured articles published to your Webflow site on a schedule, with you holding the veto.
FAQ
## GEO audit & AI visibility checker: FAQ
What is a GEO audit?
A GEO audit checks how well your site is set up to be understood, quoted, and recommended by AI answer engines like ChatGPT, Perplexity, and Google's AI. It flags the structure, schema, and content gaps that keep you out of AI answers.
What is an AI visibility checker?
An AI visibility checker tests whether AI engines can find and cite your content. HarperFlow's free audit works as one: it reviews your page against the signals AI engines reward and shows what to fix.
Is the GEO audit really free?
Yes. Enter your URL and we email you the report, at no charge and with no card required.
What does the audit check?
It looks at answer-first structure, schema, crawlability for AI bots, citations and sourcing, and other on-page signals that make content easy for AI to quote.
How long does the audit take?
The audit runs in minutes and the report is delivered to your inbox.
## Find out what AI *really* says about you.
The audit takes seconds and the report is free. If you'd rather skip straight to publishing content built to be cited, start a free trial — three full articles on your own site.
[Run my free audit ↑](#audit-widget) [Start free today →](https://app.harperflow.io/auth/register)
Keep exploring
## Related reading
[Research
The 2026 State of AI Citations
Three engines, the same buyer questions — who AI recommends, measured. Read →](https://www.harperflow.io/state-of-ai-citations) [Research
Can your site get cited?
The one factor that survived testing: structure, at p=0.004. Read →](https://www.harperflow.io/can-your-site-get-cited-by-ai) [Product
How HarperFlow works
Discover, draft, schedule, veto, publish — the autopilot loop end to end. Read →](https://www.harperflow.io/features)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/citation-run
# One keyword in. A source-backed article out.
A real pipeline run on our own blog. Every section, source, and figure below is what HarperFlow produced and published — and every source links to the page it came from, so you can check all of it.
Target keyword `aeo webflow`
The run
[AEO on Webflow: How to Get Cited Without Enterprise](https://www.harperflow.io/harperflow-blog/aeo-on-webflow-how-to-get-cited-without-enterprise)
harperflow.io · Sep 13, 2026
**2,455** words**7** sources**8** FAQs
[Read the live article](https://www.harperflow.io/harperflow-blog/aeo-on-webflow-how-to-get-cited-without-enterprise)
Onboard
## Connect the site HarperFlow publishes to.
This run used our own Webflow site. HarperFlow reads the blog to learn what it already covers, then publishes new articles straight into its CMS.
harperflow.io/harperflow-blog
HarperFlow Blog
Source-backed articles, published to the CMS
How to Add GTM to Webflow (and Verify It Actually Fires)
How to Set Up a Custom 404 Page in Webflow
AEO on Webflow: How to Get Cited Without Enterprise
Connection
Webflow site connected
Publishes to
HarperFlow Blog collection
Articles in the collection
32
This run
Sep 13, 2026
Discover
## Find a question worth answering.
The pipeline picks a target keyword and works out the gap behind it from what the sources already say.
Target keyword
aeo webflow
A reader question it answers: **“Can I use Webflow’s native AEO agents if I’m on a CMS or Business plan?”**
> Webflow’s own AEO is a "native, closed-loop solution for enterprise marketing teams", and its AEO agents "require a Team or Enterprise Platform plan".
>
> [webflow.com/feature/aeo](https://webflow.com/feature/aeo) [help.webflow.com](https://help.webflow.com/hc/en-us/articles/51704193319955-AEO-agents-overview)
>
> The gap: teams on other paid plans still need AEO. So the article shows how to do it with custom code.
Build
## The pipeline plans, researches, checks, and assembles.
These are the pipeline’s real stages, with what each one produced on this run. Press play to step through.
01
### Plan the article
An orchestrator outlines the sections and what each one has to cover.
1. What AEO on Webflow Actually Means (And What It Doesn’t)
2. Webflow Native AEO vs. Manual DIY vs. Automated Content Engines
3. Adding FAQPage, Article, and Organization Schema in Webflow’s Custom Code
4. Writing Direct-Answer Content That Earns Citations
5. Keeping Citations Current: From One-Time Fixes to a Publishing Rhythm
02
### Research each section in parallel
Each section is written by its own agent, which researches with web search and full-page reads.
7 sources across
webflow.comhelp.webflow.comcmswire.com
03
### Check the facts
Figures are re-read against their source pages and links are verified before anything publishes. On this run, two Webflow pages described the same gate differently:
"all customers on an Enterprise **Site** plan get these agents" [webflow.com/feature/aeo](https://webflow.com/feature/aeo)
"agents require a Team or Enterprise **Platform** plan" [help.webflow.com](https://help.webflow.com/hc/en-us/articles/51704193319955-AEO-agents-overview)
Site plan and Platform plan are different things. Rather than flatten them into one claim, the article **kept both and attributed each**: "the feature page states all customers on an Enterprise Site plan get these agents … and the help docs note agents require a Team or Enterprise Platform plan with Workspace AI enabled."
04
### Assemble and QA
Sections are merged, then the FAQ, TL;DR, and metadata are written and the draft gets a final QA pass.
2,455 words · 8 FAQs · 5 key takeaways · TL;DR and metadata
Published
## Live on the blog, sources attached.
This is the article as it shipped on September 13, 2026. Every source below opens the page it came from.
harperflow.io/harperflow-blogSep 13, 20269 min read
### AEO on Webflow: How to Get Cited Without Enterprise
Webflow native AEO is a closed-loop solution for enterprise teams that launched May 21 2026 and gates AEO agents to Enterprise Site plans. Manual DIY AEO works on any paid Webflow plan using custom code embeds for FAQPage, Article, and Organization schema without Enterprise.
#### Key takeaways (3 of 5)
- Webflow custom code only supports HTML, CSS, and JS, with a 50,000 character limit for each Head and Footer section in site and page settings.
- Native AEO agents flag broken links, stale metadata, missing alt text and schema issues, and analytics tracks AI bot engagement and brand citations.
- Page-level custom code appears after site-wide code, and AI credit consumption for AEO agents started June 29 2026 under Analyze for Enterprise.
Published to harperflow.io
What shipped
Words
2,455
Sources
7
FAQ entries
8
Key takeaways
5
Sources — all 7, clickable
1. [Own your visibility in AI search](https://webflow.com/feature/aeo)webflow.com
2. [AEO agents overview](https://help.webflow.com/hc/en-us/articles/51704193319955-AEO-agents-overview)help.webflow.com
3. [Webflow AEO: Now available for Enterprise](https://webflow.com/updates/now-available-webflow-aeo)webflow.com
4. [Webflow Launches AEO for Enterprise](https://www.cmswire.com/digital-experience/webflow-launches-aeo-for-enterprise/)cmswire.com
5. [Custom code in head and body tags](https://help.webflow.com/hc/en-us/articles/33961357265299-Custom-code-in-head-and-body-tags)help.webflow.com
6. [Custom code in the CMS](https://help.webflow.com/hc/en-us/articles/33961236623635-Custom-code-in-the-CMS)help.webflow.com
7. [AEO with Webflow](https://webflow.com/solutions/aeo)webflow.com
[Read the live article](https://www.harperflow.io/harperflow-blog/aeo-on-webflow-how-to-get-cited-without-enterprise)
This ran on our blog. Run it on yours.
7-day trial · 3 articles · card required
[Start your trial](https://app.harperflow.io/auth/register)
Back Play the run
Real run from [harperflow.io/harperflow-blog/aeo-on-webflow-how-to-get-cited-without-enterprise](https://www.harperflow.io/harperflow-blog/aeo-on-webflow-how-to-get-cited-without-enterprise), published September 13, 2026. Section titles, sources, quoted claims, and figures are shown as published.
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/how-to-get-cited-by-ai
# How to Get *Cited* by AI Answer Engines
AI answers now name a shortlist of companies, not ten blue links. Here is what ChatGPT, Perplexity, and Google's AI actually quote — and the seven moves that put your brand in the answer.
[Start free→](https://app.harperflow.io/auth/register?credits=10) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Built on our 2026 State of AI Citations study — **117 real AI answers** across 39 buyer questions and 3 engines.
The short answer
## How do you get cited by AI?
TL;DR
To get cited by AI, publish a page that answers the exact question up top, backs it with at least one original data point, and is structured for extraction — clear headings, a table, and an FAQ. Then earn a presence in the sources each engine already trusts. AI quotes what is easy to lift and hard to dispute; these moves make your page both.
### What "cited by AI" actually means
Being cited by AI means an answer engine — ChatGPT, Perplexity, or Google's AI Overviews — names your brand or quotes your page inside its answer. It is the AI-era version of ranking, except the engine returns a few names, not a page of links.
### Why it is not the same as SEO
Classic SEO competes for a link on a results page. Getting cited competes to be the sentence the model repeats. Structure, original data, and being present in trusted sources matter more than keyword density — and the payoff is being the recommendation, not option seven.
Why now
## The shortlist replaced the search page
When a buyer asks an AI engine for a recommendation, it returns a handful of names and stops. There is no page two to climb to.
### There are no ten blue links
A generative answer names a few companies and moves on. About 98% of commercial AI answers name specific companies by name. If you are not in that shortlist, the buyer never sees you — and never knows you were an option.
### Three engines, three different answers
In our study, all three engines agreed on a single top pick just 56% of the time, and their full lists of names overlapped only 7%. Getting cited is not one race — it is three, and you win them separately. [See the research](https://www.harperflow.io/geo-research).
### The window is open
Most sites still write for Google's ten links, not for what a model will quote. That gap is the opportunity: the pages built to be cited today are the ones answer engines will lean on tomorrow.
The evidence
## What AI actually cites
We logged every source across 117 answers. Four patterns held up — and each one is something you can act on.
| Pattern | What we measured | What to do about it |
| --- | --- | --- |
| Listicles & "best-of" pages | \~77% of cited pages | Structure the page as a clear, comparable list — not a wall of prose. |
| Original data | only \~31% of cited pages carry any | Publish one real number and you are in the minority AI rewards. |
| Structure (FAQ + tables) | cited 75% vs 22% of the time (p=0.004) | Add an FAQ and a table — the formats AI lifts most. |
| The long tail | top three domains held \~18–21% of citations | You do not need to be #1; presence across answers compounds. |
Sources also move fast: Reddit appeared in 87% of one engine's answers in a single run, then 18% fifteen days later. Optimise for the pattern, not a single source. [Read the Reddit collapse](https://www.harperflow.io/the-reddit-collapse).
The playbook
## How to get cited by AI, step by step
Seven moves, in order. The first four earn the citation; the last three compound it.
### 1. Publish the format AI quotes
Build structured best-of, comparison, and direct-answer pages. Around 77% of cited pages are listicles or best-of guides — give the engine the shape it already prefers to quote.
### 2. Add original data
A benchmark, a survey, a number only you have. Only \~31% of cited pages carry any data, so a single real figure sets your page apart and gives the model something quotable.
### 3. Answer the exact question, up top
Put a two-to-four-sentence answer in the first screen, phrased the way people ask. That self-contained passage is the block an engine lifts verbatim.
### 4. Be present where the engine looks
Engines cite the sources they trust — Reddit, YouTube, industry sites, and review platforms. Show up there honestly so your name travels into the answer from somewhere the model already reads.
### 5. Structure for extraction
Use clear headings, tables, FAQs, and short definitional sentences. In our data, structured pages were cited far more often than plain prose (75% vs 22%).
### 6. Cite primary sources, with dates
Link authoritative references and date your own claims. Verifiable, current pages are easier for a machine to trust — and to repeat without hedging.
### 7. Track it across all three engines
They disagree more than they agree, so measure ChatGPT, Perplexity, and Google separately and fix the gaps. [A free GEO audit](https://www.harperflow.io/geo-website-audit) shows where you stand today.
The shortcut
## Or run your blog on autopilot
Doing all seven, every week, is the hard part. HarperFlow is a publishing pipeline that researches, writes, cites, and publishes GEO articles built to be cited — straight to your Webflow blog. Same moves, on autopilot. See [how the pipeline works](https://www.harperflow.io/features), or read the [Webflow-specific playbook](https://www.harperflow.io/how-to-get-your-webflow-site-cited-by-ai).
## Only about *31%* of cited pages carry any original data.
Publish one real number and you are already ahead of two-thirds of the web.
Avoid this
## Where most sites go wrong
### Thin, generic AI content
Publishing volume with no data or point of view gives the model nothing to quote. Cheap content is not cheap — it is spend with no citation to show for it.
### No structure to lift
A great argument buried in unbroken prose will not get extracted. Without a clear answer, a table, or an FAQ, the engine skips you for a page it can quote cleanly.
### Optimising for one engine
Winning in ChatGPT says little about Perplexity or Google — they overlap only 7%. Treat each as its own channel or you will be invisible on two of three.
Keep reading
## The research behind this guide
Every number here comes from our own study. Dig into [the GEO research hub](https://www.harperflow.io/geo-research), the full [2026 State of AI Citations](https://www.harperflow.io/state-of-ai-citations), [whether your own site can get cited](https://www.harperflow.io/can-your-site-get-cited-by-ai), and the definitional guide to [generative engine optimization](https://www.harperflow.io/generative-engine-optimization).
FAQ
## Getting cited by AI: common questions
What does it mean to be "cited by AI"?
It means an answer engine like ChatGPT, Perplexity, or Google's AI Overviews names your brand or quotes your page in its answer. Instead of ranking on a results page, you become one of the few sources the engine repeats.
Is getting cited by AI different from SEO?
They overlap but are not the same. Classic SEO competes for links on a results page; getting cited (GEO) competes to be quoted inside a generated answer. Structure, original data, and presence in trusted sources matter more than keyword density.
Can you guarantee my site will get cited?
No — and be wary of anyone who promises it. Citations depend on the engines, your authority, and time. What you can do is remove every technical and structural reason a page would not be cited, which is what this guide covers.
Which pages does AI cite most?
In our study, roughly 77% of cited pages were listicles or best-of guides, and structured pages with FAQs and tables were cited far more often than plain prose. Original-data pages punch above their weight.
How long does it take to get cited by AI?
There is no fixed timeline. Answer engines re-crawl and re-rank continuously, and citations can shift week to week — one source in our study went from 87% to 18% in fifteen days. Treat it as ongoing, not one-and-done.
Do I have to be on Reddit to get cited?
No. Reddit is heavily cited by some engines, but its share is volatile and varies by engine. Being present in the sources your buyers' engine trusts helps; relying on any single source does not.
Keep reading
## Related guides
[LLM SEO](https://www.harperflow.io/llm-seo)· [free GEO audit](https://www.harperflow.io/geo-website-audit)
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
## See where AI stands on your site
Run a free GEO audit and see exactly where your pages can become easier for AI to quote and recommend.
[Run a free GEO audit→](https://www.harperflow.io/geo-website-audit) [Start free](https://app.harperflow.io/auth/register?credits=10)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/how-to-get-your-webflow-site-cited-by-ai
# How to Get Your *Webflow* Site Cited by AI
Webflow gives you a fast, clean site. Getting it quoted by ChatGPT, Perplexity, and Google's AI takes a few specific moves — CMS structure, schema, crawler access, and content built to be cited. Here is the Webflow-specific playbook.
[Start free→](https://app.harperflow.io/auth/register?credits=10) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
From the team that publishes GEO articles straight into Webflow — **built to be cited**, not just posted.
The short answer
## How do you get a Webflow site cited by AI?
TL;DR
Structure your Webflow CMS content answer-first, add JSON-LD schema through Webflow's per-page custom code, let AI crawlers in via robots.txt (and add an llms.txt), interlink your Collection items, and publish on a steady cadence. Webflow gives you every control you need; the content is what earns the citation.
### Webflow is a strong base, not the finish line
Webflow ships clean, fast, semantic HTML that AI crawlers read easily. But a fast site is not a cited site — you still need the answer-first structure and the presence that answer engines reward.
### What "cited" means here
When someone asks ChatGPT or Perplexity a question in your niche, your Webflow page is named or quoted in the answer. That is the goal — becoming the recommendation, not just ranking on Google.
Why now
## Why Webflow sites need a GEO layer
AI answers return a shortlist, not ten links — and the engines disagree with each other.
### The shortlist is the new results page
About 98% of commercial AI answers name specific companies. If your Webflow site is not on that short list, the buyer never sees it — there is no page two to climb.
### Three engines, three answers
In our study, all three engines agreed on a single top pick just 56% of the time, and their full lists overlapped only 7%. Optimise for all three, not one. [See the research](https://www.harperflow.io/geo-research).
### Webflow will not do it for you
Webflow handles hosting, speed, and clean markup. It does not write answer-first content, plan your schema, or keep you publishing — that part is on you, or on a pipeline that does it for you.
The playbook
## How to get your Webflow site cited, step by step
Six moves, all Webflow-native.
### 1. Structure CMS content answer-first
In your Collection template, lead every article with a two-to-four-sentence direct answer — a dedicated summary or Rich Text field. That self-contained passage is what an engine lifts.
### 2. Add JSON-LD schema in Webflow
Use Page Settings → Custom Code (or the page head) to add Article, FAQPage, and HowTo JSON-LD. Webflow supports per-page custom code on paid site plans; validate with Google's Rich Results Test.
### 3. Let AI crawlers in
In Site Settings → SEO, edit robots.txt to allow GPTBot, PerplexityBot, Google-Extended, and ClaudeBot. Add an llms.txt pointing crawlers at your best pages — Webflow has no native slot, so host it via a redirect or a Cloudflare Worker.
### 4. Interlink your Collections
Link related Collection items and hub pages with descriptive anchor text. Internal links spread authority and help engines and readers find your best answers.
### 5. Publish on a cadence
Answer engines re-crawl continuously and citations drift — one source in our study fell from 87% to 18% in fifteen days. Consistent publishing keeps you in a shifting index. [See the drift](https://www.harperflow.io/the-reddit-collapse).
### 6. Track it across engines
Check ChatGPT, Perplexity, and Google separately, because they disagree. [A free GEO audit](https://www.harperflow.io/geo-website-audit) shows where your Webflow site stands.
The checklist
## The Webflow GEO checklist
Every setup item, where it lives in Webflow, and why it earns citations.
| Setup step | Where in Webflow | Why it matters |
| --- | --- | --- |
| Answer-first summary | CMS Collection field + template | The extractable passage AI quotes |
| JSON-LD schema | Page Settings → Custom Code | Machine-readable context for engines |
| robots.txt (allow AI bots) | Site Settings → SEO | Lets GPTBot, PerplexityBot and others read you |
| llms.txt | Hosted file / redirect (no native slot) | Points crawlers at your best pages |
| Internal links | Collection references + Rich Text | Spreads authority; aids discovery |
| Publishing cadence | Webflow CMS / scheduled publish | Keeps you in a shifting index |
| Alt text + clean headings | Asset settings + H2/H3 | Accessible, extractable structure |
The shortcut
## Or publish to Webflow on autopilot
HarperFlow does exactly this, for you. It researches, writes, cites, and publishes GEO articles built to be cited — straight into your Webflow CMS, on a schedule, with a human veto before anything goes live. Same playbook, no manual steps. See [how the pipeline works](https://www.harperflow.io/features) or the [fundamentals of getting cited](https://www.harperflow.io/how-to-get-cited-by-ai).
## Webflow gives you the controls. *Content* earns the citation.
The pages AI quotes are not the prettiest — they are the clearest, best-structured, best-sourced answers.
FAQ
## Webflow and AI search: common questions
Is Webflow good for AI SEO and GEO?
Yes. Webflow outputs clean, fast, semantic HTML that AI crawlers read easily, and it supports the per-page schema and robots.txt control GEO needs. The platform is a strong base; the content and structure are what earn citations.
How do I add schema markup in Webflow?
Use Page Settings → Custom Code to paste JSON-LD (Article, FAQPage, HowTo) into the page head, or add it site-wide in Project Settings. Validate with Google's Rich Results Test before publishing.
Can I add an llms.txt file to a Webflow site?
Not natively — Webflow only exposes robots.txt among root files. Host llms.txt via a redirect or a Cloudflare Worker in front of your domain, pointing AI crawlers at your best pages.
Does Webflow block AI crawlers by default?
No, but you control it in Site Settings → SEO → robots.txt. Make sure GPTBot, PerplexityBot, Google-Extended, and ClaudeBot are allowed if you want to be cited.
How long until my Webflow site gets cited by AI?
There is no fixed timeline. Engines re-crawl continuously and citations shift week to week, so publish consistently and track it as an ongoing effort.
Do I need to leave Webflow to do GEO?
No. Every move here works within Webflow. A pipeline like HarperFlow automates it end to end, but the steps themselves are Webflow-native.
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
## See how AI reads your Webflow site
Run a free GEO audit on your Webflow URL and see exactly what to fix first.
[Run a free GEO audit→](https://www.harperflow.io/geo-website-audit) [Start free](https://app.harperflow.io/auth/register?credits=10)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/how-to-rank-on-chatgpt
# How to Rank on *ChatGPT* (Get Recommended in AI Answers)
ChatGPT does not show ten blue links — it names a few brands and moves on. Here is how to become one of the names it recommends, with the exact on-page and off-page moves, backed by 117 tested answers.
[Start free→](https://app.harperflow.io/auth/register?credits=10) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Backed by our 2026 State of AI Citations study — **117 answers** across ChatGPT, Perplexity and Gemini.
The short answer
## How do you rank on ChatGPT?
TL;DR
To rank on ChatGPT, make your brand the easy, safe answer: publish pages that answer the exact question up top, back them with original data, and be present in the sources ChatGPT already trusts. Then structure everything for extraction. ChatGPT repeats what is quotable and well-sourced — give it that.
### Ranking on ChatGPT is really being cited
ChatGPT does not rank pages the way Google does. It composes an answer and names sources. Ranking means becoming one of the brands or pages it repeats — something you influence, not a position you buy.
### It is not classic SEO
Keyword density and backlinks help less than a clear, structured, sourced answer a model can lift with confidence. The bar is quotability, not just relevance.
Why now
## There is no page two in ChatGPT
A generative answer is a shortlist, not a results page.
### One answer, a few names
About 98% of commercial AI answers name specific companies. ChatGPT picks a handful and stops — if you are not among them, the buyer never sees you.
### Quotable beats optimised
ChatGPT favours sources it can quote without hedging: clear claims, real data, credible citations. Write to be repeated, not just ranked.
## About *98%* of commercial AI answers name specific companies.
The only question is whether ChatGPT names you.
The playbook
## How to rank on ChatGPT, step by step
Six moves that make your brand the safe answer.
### 1. Be quotable, up top
Put a two-to-four-sentence answer to the exact question in the first screen. That self-contained passage is what ChatGPT lifts into its reply.
### 2. Add original data
A number only you have makes your page the citation, not a rewrite of everyone else's. Only \~31% of cited pages carry any data.
### 3. Be present where ChatGPT looks
It leans on trusted sources — industry sites, Reddit, YouTube, documentation. Show up there honestly so your name reaches the answer.
### 4. Structure for extraction
Clear headings, tables, FAQs, and short definitional sentences. Structured pages were cited far more often than plain prose (75% vs 22%).
### 5. Cite sources, with dates
Verifiable, current claims are safer for a model to repeat. Link authoritative references and date your own numbers.
### 6. Do not stop at ChatGPT
It agrees with Perplexity and Gemini only 56% of the time. Win each engine on its own terms. [See how differently they answer](https://www.harperflow.io/one-engines-favourite).
Three channels
## ChatGPT vs Perplexity vs Gemini
They answer the same question differently, so treat each as its own channel.
| Engine | How it answers | What it rewards |
| --- | --- | --- |
| ChatGPT | Conversational shortlist; cites sources when browsing | Clear, quotable, well-sourced answers |
| Perplexity | Citation-heavy; surfaces many sources per answer | Presence across trusted sources; volatile week to week |
| Gemini | Blends Google's index with AI Overviews | Traditional SEO signals plus structured, authoritative pages |
In our study the three agreed on a single top pick just 56% of the time, and their full lists overlapped only 7%. [See the research](https://www.harperflow.io/geo-research).
The shortcut
## Or make every article quotable, automatically
HarperFlow researches, writes, cites, and publishes GEO articles built to be cited — straight to your Webflow blog, on a schedule. The same moves that earn a ChatGPT mention, without doing them by hand. See [how it works](https://www.harperflow.io/features) or the [fundamentals of getting cited](https://www.harperflow.io/how-to-get-cited-by-ai).
FAQ
## Ranking on ChatGPT: common questions
Can you really "rank" on ChatGPT?
Not like Google. ChatGPT does not rank pages; it composes an answer and names sources. Ranking means becoming one of the brands or pages it repeats, which you influence with structure, data, and presence.
How is ranking on ChatGPT different from SEO?
SEO competes for links on a results page; ChatGPT competes to be quoted in an answer. Clear answers, original data, and credible sources matter more than keyword density.
Does ChatGPT use Google or Bing?
When browsing, ChatGPT search draws on the live web and its own model knowledge. Being crawlable and citable matters, and so does being present in the sources it trusts.
What is "ChatGPT SEO"?
It is optimising to appear in ChatGPT's answers rather than Google's links, a subset of generative engine optimization (GEO). The tactics overlap with SEO but prioritise extractable, well-sourced content.
How do I get ChatGPT to recommend my brand?
Publish structured, original, well-sourced pages that answer buyer questions, and build an honest presence in the sources ChatGPT cites. There are no guarantees, but these remove the reasons it would skip you.
How long does it take?
There is no fixed timeline; ChatGPT's browsing index and training data change over time. Treat it as ongoing and track your presence across engines.
Keep reading
## Related guides
[LLM SEO](https://www.harperflow.io/llm-seo)· [free GEO audit](https://www.harperflow.io/geo-website-audit)
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
## See if ChatGPT can quote your site
Run a free GEO audit and see what to fix so AI answers can find, quote, and recommend you.
[Run a free GEO audit→](https://www.harperflow.io/geo-website-audit) [Start free](https://app.harperflow.io/auth/register?credits=10)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/generative-engine-optimization
# Generative Engine Optimization: The Complete 2026 Guide
AI answer engines now sit between your buyers and your website — and a growing share of searches end in the answer, not a click. Generative engine optimization (GEO) is how your content becomes the source those engines quote. This guide covers what GEO is, how it differs from SEO, how engines decide what to cite, and how to earn those citations.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [Start free today →](https://app.harperflow.io/auth/register)
Researched and updated **August 2026**. Grounded in our own AI-citation research.
The honest answer
## What is generative engine optimization?
The short answer
Generative engine optimization (GEO) is the practice of creating and structuring content so AI answer engines — ChatGPT, Perplexity, Google's AI Overviews, Gemini — quote and cite it in their answers. Where classic SEO competes for a ranked link on a results page, GEO competes to be the source the model paraphrases when it answers the question directly. Same goal, be found; new surface, the answer instead of the link.
### Why it matters now
AI answers increasingly sit between a searcher and your site, and a growing share of research ends in the answer with no click at all. GEO is how you stay visible when the click disappears — by becoming the source the machine trusts enough to quote.
### GEO, AEO, LLMO — same shift
Answer Engine Optimization (AEO) and LLM Optimization (LLMO) describe the same move under different names. The tactics converge on one idea: structured, sourced, genuinely useful content that a language model can lift with confidence.
GEO vs SEO
## What actually changes from SEO to GEO
GEO doesn't replace SEO — it extends it to a new surface. But three things genuinely change.
### The unit of victory
SEO wins a ranked position; GEO wins a citation inside the answer. You can rank third and still be the source the AI quotes — or rank first and be ignored while the model paraphrases someone else.
### What the machine rewards
Answer engines favour content that responds directly, cites primary sources, and is cleanly structured with headings, FAQs, and tables — not keyword density or backlink volume alone.
### How you measure it
SEO tracks positions and clicks. GEO tracks whether engines mention and cite you — across prompts and across engines — because a citation you can't see is one you can't grow.
The mechanics
## How AI engines choose what to cite
No one has the full algorithm, but the signals that consistently correlate with being cited are clear — and several come straight from our own testing.
- **Direct answers.** A clear, self-contained answer near the top of the page is easy to lift; a buried answer gets skipped.
- **Structure.** FAQ blocks, tables, and descriptive headings are disproportionately quoted — our research found structured pages cited about 75% of the time versus 22% for unstructured.
- **Citations and sources.** Content grounded in primary data reads as trustworthy to a model trained to avoid inventing facts.
- **Freshness.** Recently updated, dated content is preferred for anything time-sensitive.
- **Corroboration.** When several reputable sources say the same thing the same way, the odds any one of them is cited go up.
- **Entity clarity.** Consistent naming of your brand, products, and claims helps the model attach facts to you specifically.
We tested these directly — see [can your own site get cited by AI](https://www.harperflow.io/can-your-site-get-cited-by-ai) for the numbers behind the structure finding.
The playbook
## How to do GEO, step by step
A repeatable loop you can run on any page you want cited. None of it is a platform setting — it's editorial discipline.
1. Start from the questions your buyers actually ask an AI, not just keywords.
2. Answer each one in the first paragraph, in a single quotable sentence.
3. Structure the page with clear H2s, an FAQ block, and a table wherever it fits.
4. Ground every non-obvious claim in a citable, primary source.
5. Add Article, FAQ, and Breadcrumb schema so engines can parse it cleanly.
6. Publish consistently and keep your cornerstone pages fresh and dated.
7. Measure citations across ChatGPT, Perplexity, and Gemini, and feed what wins into the next piece.
Do this by hand, or run it on autopilot. Either way, GEO rewards the teams that treat it as an ongoing practice, not a one-time audit.
Tools & approach
## GEO tools, and the honest way to run it
The GEO tool market splits into a few jobs. Most tools do one; the results come from doing all of them.
### What GEO tools do
Some measure AI-search visibility — are you cited, and by which engine? Some help you draft. Some audit structure and schema. Each is useful, but a visibility score with nothing publishing behind it doesn't move the needle.
### The pipeline that actually compounds
The durable version is a loop: research a real question, write a cited and structured answer, audit it, keep a human veto, publish to your CMS, and measure the citations. That full loop — not just scoring — is what HarperFlow runs on your Webflow, WordPress, or Shopify site. [See how it works](https://www.harperflow.io/features), or [run a free GEO audit](https://www.harperflow.io/geo-website-audit) to see where you stand today.
## Search sent you a click. AI sends you a *citation* — or nothing.
GEO is how you stay the source when the answer replaces the link.
Automate the cadence
## Automating it
The one checklist item that compounds — publishing structured content on a schedule — is also the one you can hand to software instead of a retainer.
### Webflow-first autopilot
That is what HarperFlow does, and it is Webflow-first: it runs on autopilot to discover topics, draft GEO-optimized articles, schedule them, email you a veto digest, and publish the un-vetoed drafts straight to your Webflow, WordPress, or Shopify blog.
### Built to be cited
The AI-citation angle is deliberate. HarperFlow's own analysis of AI answers found that how a page is *structured* predicts whether it gets cited — a real, measured effect (p=0.004), documented in [Can your site get cited by AI](https://www.harperflow.io/can-your-site-get-cited-by-ai). Every article HarperFlow publishes is built to that structural standard, and each one carries a human veto window before it goes live, so automation never means losing control. See how the pipeline fits together in [Features](https://www.harperflow.io/features), or read the setup path in the [docs](https://www.harperflow.io/docs).
### Monthly, usage-based pricing
Pricing is monthly and usage-based: the Flex plan starts at **$62/mo for Webflow and WordPress for 5 articles** ($12.40 each) and scales up to 30, with no legacy tier lock-in. Full detail on [pricing](https://www.harperflow.io/pricing).
Put GEO to work on Webflow
## Related guides
GEO in practice, on your own site: [Webflow SEO — the complete setup](https://www.harperflow.io/webflow-seo), [Webflow AI — using AI for content without the slop](https://www.harperflow.io/webflow-ai), and [The Webflow Blog Guide](https://www.harperflow.io/webflow-blog) on building a blog that gets cited. Or see the original research behind this page in our [GEO research](https://www.harperflow.io/geo-research).
New to this? Start with the practical playbook — [how to get cited by AI](https://www.harperflow.io/how-to-get-cited-by-ai) — then use this guide for the deeper GEO mechanics.
Keep reading
## Related guides
[LLM SEO](https://www.harperflow.io/llm-seo)· [free GEO audit](https://www.harperflow.io/geo-website-audit)
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions
## Frequently asked questions
What is generative engine optimization (GEO)?
Generative engine optimization (GEO) is optimizing your content so AI answer engines like ChatGPT, Perplexity, and Google's AI Overviews quote and cite it in their answers. It's the successor to ranking a link: you compete to be the source the model paraphrases when it answers directly.
Is GEO the same as SEO?
No. SEO competes for a ranked link on a results page; GEO competes to be cited inside an AI-generated answer. They overlap — clean structure and good content help both — but the win condition, and the way you measure it, are different.
Is GEO the same as AEO or LLMO?
Effectively yes. Answer Engine Optimization (AEO) and LLM Optimization (LLMO) are near-synonyms for GEO. All three describe optimizing content to be surfaced and cited by AI answer systems, and the tactics are the same.
Does GEO replace SEO?
No — it extends it. Traditional search isn't going away, and the fundamentals (crawlable, well-structured, useful pages) serve both. GEO adds a new surface, the AI answer, that increasingly captures the search you used to win with a ranked link.
How do I get my site cited by ChatGPT?
Answer the question directly in the first paragraph, structure the page with headings, an FAQ, and a table, cite primary sources, add schema, and keep it fresh. Structured, sourced pages are quoted far more often than walls of text.
What are the best GEO tools?
GEO tools split into visibility trackers (are you cited?), AI writers, and structure or schema auditors. The most effective setup is a pipeline that researches, writes with citations, audits, publishes, and measures — which is what HarperFlow does end to end.
How do I measure GEO?
Track whether AI engines mention and cite your brand across a set of real buyer prompts, repeated over time and across ChatGPT, Perplexity, and Gemini. A free GEO audit is the fastest way to get a first baseline.
## Run a free GEO audit
Before you hire anyone or touch a line of markup, see where your Webflow site actually stands in AI answers — no login, sourced and dated like everything else we publish.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [See how HarperFlow works](https://www.harperflow.io/features)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/llm-seo
# LLM SEOHow to Get Your Site Into AI Answers.
LLM SEO is the work of getting your pages quoted inside AI answers — the summaries Perplexity, ChatGPT, Google's AI Overviews and Claude hand back instead of ten blue links. It overlaps with classic search work but optimises for a different act: being **cited**, not just ranked. This guide covers what the term means, how large language models pick their sources, the tools people compare when they shop for an LLM SEO tool, and how to do the work without buying one. Every figure here was measured, not assumed — see the dated line below.
[Run the free GEO audit →](https://www.harperflow.io/geo-website-audit) [See Flex pricing →](https://www.harperflow.io/pricing)
Keyword and SERP data measured **27 July 2026** via Semrush (US).
The quick answer
"LLM SEO" and "LLMO" name the same job: getting your pages quoted inside AI answers, not just ranked in ten blue links. Most tools people compare are *trackers* — they tell you where AI mentions you. But getting cited is a publishing problem: structured, quotable pages, shipped on a cadence. HarperFlow sits on that side of the line — a writer-publisher, not a monitor.
Definitions
## What LLM SEO and LLMO mean
LLM SEO — often written LLMO, for large language model optimisation — is search optimisation aimed at generative engines. Traditional SEO earns a ranking; LLM SEO earns a citation: the link or named mention a model pulls into its answer. The two overlap, because a crawlable, well-structured page helps both, but the payoff differs. A page can sit on page two of Google and still be the source a model quotes, because models assemble one answer from many pages at once rather than sending a reader to a single winner.
The vocabulary is unsettled — LLM SEO, LLMO, GEO (generative engine optimisation) and plain "AI SEO" all point at roughly the same idea. We use LLM SEO because it is the term the most people actually search.
The evidence
## How LLMs choose what to cite
We measured this rather than guessed — three findings, each from a dated study.
### There is no fixed top ten to climb
Across 328 Perplexity citations for 39 buyer questions, no single domain dominated: 86% of cited domains appeared exactly once, 34 different domains held a top-citation spot, and the long tail carried about 85% of all citations. The lesson from [that study](https://www.harperflow.io/never-on-the-podium) is that models spread their sourcing wide — so being one of many credible, quotable pages is a realistic goal, not a moonshot.
### Structure predicts citation
In a companion study, pages carrying [FAQ blocks or data tables](https://www.harperflow.io/can-your-site-get-cited-by-ai) appeared in 75% of cited brand pages versus 22% of cited third-party pages — the only trait that reached statistical significance (p=0.004). Named authors, listicle formatting and raw word count did not. If you change one thing, make your key pages trivially extractable.
### Citation share is not durable
Source authority moves fast. Reddit went from appearing in 87% of Perplexity answers to 18% in [fifteen days](https://www.harperflow.io/the-reddit-collapse) — byte-identical questions, same engine — while the answers themselves actually got richer. Citation share has to be re-earned with current, structured pages. That volatility is the whole argument for treating LLM SEO as an ongoing publishing habit rather than a one-time audit.
Tools
## LLM SEO tools compared
People shop for this a dozen ways, and the phrasing tells you what they are really after. The tracking crowd searches for the best llm seo tracker, the best llm seo trackers, a best llm seo rank tracker, the best llm seo tracking software, or a best llm seo checking tool, best llm seo checking tools or best llm seo checking software to confirm where they stand. The building crowd searches for the best llm seo tool, the best llm seo software, a best llm seo analysis tool or best llm seo analysis software, or a best llm seo optimization tool and best llm seo optimization tools. Every one of those collapses into a single decision: do you want something that tells you where you are cited, or something that changes it?
Most named products are trackers, and that is genuinely useful — a good tracker watches which models mention your brand, on which prompts, and turns it into a dashboard. Dedicated monitors win on coverage: they watch more engines and prompts than we do. What a tracker does not do is write and publish the structured, citable pages the last section showed you need. HarperFlow sits on the other side of that line — it is a writer-publisher, not a monitor.
| Tool | Writes the content | Publishes to your CMS | Webflow, WordPress & Shopify | Human veto before publish | Built to do |
| --- | --- | --- | --- | --- | --- |
| HarperFlow | Yes | Yeson autopilot | Nativemapped CMS fields | Yesveto digest each cycle | Write & publish AI-citable articles |
| Profound | No | No | No | n/a | Track brand mentions in AI answers |
| Otterly | No | No | No | n/a | Monitor AI search results and prompts |
| AthenaHQ | No | No | No | n/a | Monitor brand visibility in AI answers |
| SearchAtlas | Partial | Partial | Generic | No | Broad SEO platform with AI features |
These are category-level positions, not a feature audit — competitor pricing and exact capabilities were not part of this data pull, so confirm them on each vendor's own site before relying on them. Read the table for what it is: HarperFlow and the trackers are not chasing the same job. Want your AI visibility across many engines in one place? A dedicated monitor like Profound, Otterly or AthenaHQ will show you more than we do — that is their win. Want to produce the pages that get cited, on your own Webflow, WordPress, or Shopify site, with a human veto before anything ships? That is the gap we built for.
## Every LLM SEO tool answers one question: do you want to know where you're *cited* — or change it?
Trackers tell you where you stand. HarperFlow writes and publishes the pages that move it.
No tool required
## Doing it without a tool
You do not need software to start. The research points at a short, unglamorous checklist.
01
### Add an FAQ block and a data table
Give your important pages an FAQ block and at least one data table. That is the single structural trait that separated cited pages from uncited ones (p=0.004).
02
### Keep pages tight, not long
Cited pages in our sample ran *shorter* than the third-party pages they beat, so length is not the lever; extractability is. HarperFlow's own [citation research](https://www.harperflow.io/state-of-ai-citations) found the same structural bias toward clear headings, sectioned answers and data a model can lift out.
03
### Keep publishing
Because citation share decays, a page cited today can drop out within weeks. Cadence beats a one-off push.
04
### Answer real buyer questions in plain language
One question per section, so a model can quote a clean, self-contained passage.
If you would rather hand it off, an llm seo agency can run this for you — but confirm they publish structured pages to your own CMS on a cadence, not just deliver a one-time audit. That publishing cadence is exactly the gap HarperFlow automates: it discovers questions worth answering, drafts structured articles, schedules them, emails a veto digest so a human can pull anything before it goes live, then publishes what is left to your Webflow, WordPress, or Shopify blog.
Want the step-by-step version? See [how to get cited by AI](https://www.harperflow.io/how-to-get-cited-by-ai) and [how to rank on ChatGPT](https://www.harperflow.io/how-to-rank-on-chatgpt).
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions
## Frequently asked questions
What is the best LLM SEO tool?
It depends on the job. To monitor where AI answers mention you, a dedicated tracker such as Profound, Otterly or AthenaHQ will cover more engines than we do. To produce and publish the structured pages that get cited — on your own Webflow, WordPress, or Shopify site, with a human veto — that is what HarperFlow is built for. The two solve different halves of the same problem.
Do I need an LLM SEO tracker, or a tool that publishes?
A tracker tells you where you stand; it does not change it. If your pages already carry FAQ blocks and data tables and you publish regularly, a tracker is a reasonable next buy. If they do not, spend first on producing citable pages — the measurement only matters once there is something to measure.
Is LLM SEO the same as GEO or LLMO?
Close enough to treat as synonyms in practice. LLMO (large language model optimisation) and GEO (generative engine optimisation) both describe optimising for AI answers rather than blue-link rankings. We use "LLM SEO" because it is the highest-volume term people actually search.
Do I need an LLM SEO agency to do this?
No — the core checklist (structured pages, plain-language answers, a steady cadence) is doable in-house. An llm seo agency helps if you lack the time to publish consistently, but verify that they ship structured content to your CMS on a cadence, not a slide-deck audit. HarperFlow automates that same loop if you would rather not hire out.
## See where you stand — free.
Before you buy any tool, find out whether AI engines cite you at all. HarperFlow's [free GEO audit](https://www.harperflow.io/geo-website-audit) checks your site's AI visibility and shows the structural gaps — missing FAQ blocks, thin tables, pages a model cannot cleanly quote — that keep you out of AI answers. Run it on your own site; your existing content stays where it is. See [Flex pricing](https://www.harperflow.io/pricing) for the publishing engine that fixes what the audit finds — from $62/mo for Webflow and WordPress for five articles a month. For a sourced, side-by-side against a specific AI-content tool, our [Soro alternative breakdown](https://www.harperflow.io/soro-alternative) shows the format we hold ourselves to.
[Run the free GEO audit →](https://www.harperflow.io/geo-website-audit) [See Flex pricing](https://www.harperflow.io/pricing)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/ai-search-optimization-tools
# AI Visibility Tools: What They Do and Which to Use
When someone asks ChatGPT, Perplexity, or Google's AI a buying question, does your brand get named? Most AI visibility tools measure that. A smaller number help you change the answer. This honest roundup names real competitors, says where each genuinely wins, and separates the tools that track from the ones that help you earn the citation.
[Get your free GEO audit →](https://www.harperflow.io/geo-website-audit) [See every tool compared →](https://www.harperflow.io/compare)
Keyword and SERP data measured **27 July 2026** via Semrush (US).
The distinction that matters
## Trackers vs. earners
Most AI visibility tools only measure whether AI mentions you. A few help you change the answer. Know which job you are buying.
### Trackers measure
Profound, Otterly, Peec AI, AthenaHQ and the rest monitor how often ChatGPT, Perplexity and Google AI name you, and how you compare to rivals. Essential for knowing where you stand — but a tracker never changes the answer.
### Earners change the answer
Only \~31% of cited pages carry any original data, and structured pages get cited far more than plain prose. Earning a citation means publishing that content — which is what a content engine like HarperFlow does, straight to your Webflow, WordPress, or Shopify blog. Measuring is not earning. [See how citations are won](https://www.harperflow.io/how-to-get-cited-by-ai), or the [GEO tools for Webflow](https://www.harperflow.io/best-geo-tools-for-webflow).
The quick answer
AI search optimization tools do one of two jobs: *measure* whether AI answers name your brand, or *change* that by publishing citable content. Most only measure. There's no single winner — there's a winner per goal.
Pick a monitor ifYou only need to measure the AI visibility your brand earns — Profound, Otterly, and AthenaHQ are built for exactly that.
Pick HarperFlow ifYou also want to act on what you find — drafting GEO-optimized articles, scheduling them, vetoing each one, then publishing to your Webflow, WordPress, or Shopify blog.
The two jobs
## What AI search optimization tools actually do
Strip away the marketing and AI search optimization tools do one of two jobs, sometimes both.
### Job one: measurement
These tools run your buyers' questions through the major assistants on a schedule and report whether you were mentioned, which sources got cited, and how that changes over time. When people ask *what are the best AI search monitoring tools*, this is the category they mean — a scoreboard for AI answers.
### Job two: action
The second job is changing what's published on your site so models have something citable to pull from. Far fewer tools do this, because it means writing and shipping content, not just watching a dashboard.
The honest split matters. A monitoring tool tells you that you're invisible in AI answers. It does not make you visible. Knowing which job you actually need is the first decision, and it's the one most buyers skip.
The levers
## How to improve brand visibility in AI search engines
Once you know where you stand, the levers for *how to improve brand visibility in AI search engines* are narrower than traditional SEO. Models cite sources that answer a specific question clearly, in one place, on a page they can parse. That rewards three things:
01
### Publish direct answers
Answer the exact questions your buyers ask — plainly, on a page dedicated to that question.
02
### Structure for lift
Structure those answers so a model can lift them cleanly — self-contained, snippet-sized, unambiguous.
03
### Earn trusted references
Be referenced by other sources the model already trusts, so your answer arrives pre-credentialed.
Backlink volume and keyword density — the traditional SEO diet — matter far less to an answer engine than being the clearest, best-structured source on a narrow question.
What actually works
## What strategies actually work
Most advice about *what strategies improve brand visibility in AI search engines* is asserted, not measured. HarperFlow ran the measurement.
Our research found that how a page is structured predicts whether AI answers cite it, at a statistically significant level (p=0.004). The full method, dataset, and the p=0.004 result are in [can your site get cited by AI](https://www.harperflow.io/can-your-site-get-cited-by-ai); the broader citation landscape is in [the state of AI citations](https://www.harperflow.io/state-of-ai-citations).
The short version: clear question-and-answer structure, self-contained sections, and specific, dated claims outperform long, undifferentiated prose. That is a finding, not a hunch — and it's the strategy worth acting on before you buy any tool.
## Page structure predicts AI citation at *p=0.004*.
That's a finding, not a hunch — and it's the strategy worth acting on before you buy any tool.
The tools, compared
## HarperFlow, Profound, SearchAtlas, Otterly, AthenaHQ
Search "ai visibility tool," "ai visibility platform," or "ai search visibility tool" and you'll get roughly the same shortlist. Here is where each one actually fits.
| Tool | Primary job | Publishes to your site | Webflow, WordPress & Shopify | Free entry point | Where it genuinely wins |
| --- | --- | --- | --- | --- | --- |
| **HarperFlow** | Publish + optimize | YesDrafts on autopilot, you veto, it publishes | Yes | Free GEO audit; Flex from $62/mo for Webflow and WordPress | Closing the visibility gap, not just measuring it; Webflow-first |
| **Profound** | Monitor AI visibility | NoMeasurement by design | No | Demo / enterprise | Deep enterprise answer-engine analytics |
| **SearchAtlas** | Broad SEO suite + AI content | PartialContent tools; you publish | No | Trial | A full traditional-SEO toolkit in one place |
| **Otterly** | Monitor AI search mentions | NoMeasurement by design | No | Lower-cost monitoring | Simple, affordable mention tracking |
| **AthenaHQ** | Monitor AI visibility (GEO) | NoMeasurement by design | No | Demo | Purpose-built GEO monitoring dashboards |
A note on fairness: three of these are monitoring tools by design, and that is not a flaw. Profound, Otterly, and AthenaHQ are built to measure the ai visibility your brand earns in AI answers, and they do that job well. If measurement is all you want, they win on focus. SearchAtlas wins if you want one broad, traditional-SEO suite rather than a GEO specialist. HarperFlow's difference is narrow and specific: its focus is publishing — drafting GEO-optimized articles, scheduling them, letting you veto, then shipping to your Webflow, WordPress, or Shopify blog.
Product features and pricing change. The keyword and SERP data on this page is Semrush-dated; the capabilities above are category positions, so verify each vendor's current feature set and price directly before buying. HarperFlow's own entry price is read from our live [pricing page](https://www.harperflow.io/pricing): Flex starts at $62/mo for Webflow and WordPress for 5 articles ($12.40 each), billed monthly and scaling up to 30. For our full, sourced, tool-by-tool breakdowns, see the [compare page](https://www.harperflow.io/compare).
Tracking mentions
## How to track brand mentions in AI search
The best ways to track brand mentions in AI search come down to a repeatable routine, not a one-time search. Pick the questions a buyer would actually ask, run them through ChatGPT, Perplexity, Gemini, and Google's AI Overviews, and record two things: whether you were mentioned, and which sources the answer cited.
That second column is the one teams forget. AI answers are assembled from sources, and those sources move. Our [Reddit collapse](https://www.harperflow.io/the-reddit-collapse) study measured Reddit falling from 87% to 18% of Perplexity citations in fifteen days. If you only track mentions of your own name, a shift that large is invisible until your traffic has already moved. Track the sources too.
Monitor vs check
## How to monitor and check AI search visibility
Two related questions, two different answers.
### How to check AI search visibility
The quick version: ask the major assistants your handful of highest-stakes questions and note whether you're named, linked, or absent. Five minutes, no tooling, useful today.
### How to monitor AI search visibility
The ongoing version: run the same prompt set, against the same assistants, on the same cadence, and track how your share of citations moves. The value is in the consistency — change the prompts every week and you're measuring noise, not progress. This is the routine a monitoring tool automates, and it's what turns a one-off check into a trend you can act on.
Keep exploring
## Related tools and guides
[Surfer SEO alternative](https://www.harperflow.io/surferseo-alternative)· [Writesonic alternative](https://www.harperflow.io/writesonic-alternative)· [free GEO audit](https://www.harperflow.io/geo-website-audit)
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions
## Frequently asked questions
What are the best AI search monitoring tools?
It depends on what you'll do with the data. For pure measurement, Profound, Otterly, and AthenaHQ are all built for it and differ mainly on depth and price. If you also want to act on what you find — publishing content that closes the gaps a monitor surfaces — HarperFlow covers both jobs. There is no single winner; there's a winner per goal.
How do you improve brand visibility in AI search engines?
Publish clear, well-structured answers to the exact questions your buyers ask, make each answer self-contained so a model can lift it, and earn references from sources the model already trusts. Traditional link and keyword tactics help far less than being the clearest source on a narrow question.
What strategies improve brand visibility in AI search engines?
The one we can back with data is structure. Our research found page structure predicts AI citation at p=0.004 — see can your site get cited by AI. Question-and-answer formatting, self-contained sections, and specific, dated claims outperform long, undifferentiated prose.
What are the best ways to track brand mentions in AI search?
Run a fixed set of buyer questions through the major assistants on a fixed cadence, and log not just whether you're mentioned but which sources get cited. Citation sources shift fast — our Reddit collapse study measured Reddit falling from 87% to 18% of Perplexity citations in fifteen days — so track the sources, not only the mentions.
How do you monitor AI search visibility?
Fix three things and hold them steady: a prompt set (your top buyer questions), a set of assistants, and a cadence (weekly or monthly). Then watch your share of citations move over time. A tool automates the runs; the discipline is keeping the inputs constant.
How do you check AI search visibility?
For a fast read, ask the assistants your most important questions yourself and see whether you're named and linked. For a grounded baseline across your whole site, run a free GEO audit.
## Start with a free GEO audit
Before you pay for any tool on this page, see exactly where you stand. HarperFlow's [free GEO audit](https://www.harperflow.io/geo-website-audit) measures how your site shows up in AI answers today and shows the structural gaps worth fixing first — no card, no sales call. Start free today.
[Run my free GEO audit →](https://www.harperflow.io/geo-website-audit) [Compare every tool](https://www.harperflow.io/compare)
Keep exploring
## Related reading
[Research
The state of AI citations
The broader citation landscape — who gets cited in AI answers, and why. Read →](https://www.harperflow.io/state-of-ai-citations) [Method
Can your site get cited by AI?
The full dataset and the p=0.004 result behind the structure finding. Read →](https://www.harperflow.io/can-your-site-get-cited-by-ai) [Study
The Reddit collapse
Reddit fell from 87% to 18% of Perplexity citations in fifteen days. Read →](https://www.harperflow.io/the-reddit-collapse)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/best-geo-tools-for-webflow
# The Best *GEO Tools* for Webflow (2026)
Most GEO and AI-visibility tools measure whether AI mentions you. A Webflow site usually needs two things: a way to track that, and a way to create content built to be cited. Here are the tools worth your time, by job — reviewed honestly, including ours.
[Start free→](https://app.harperflow.io/auth/register?credits=10) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Reviewed by the team behind HarperFlow — we include our own tool, and say exactly where it fits (and where it does not).
The short answer
## What are the best GEO tools for a Webflow site?
TL;DR
Pick by job. To measure your AI visibility, use a tracker like Profound, Otterly, or Peec AI. To audit and optimise, SE Ranking or Semrush. To actually create and publish content built to be cited — straight into Webflow — that is HarperFlow. Most teams need one tracker and one way to produce cited content.
### What counts as a GEO tool
Generative engine optimization (GEO) tools help your brand show up in AI answers. They fall into three jobs: tracking your visibility, auditing your pages, and creating content that earns citations.
### Why "for Webflow" matters
Most trackers are platform-agnostic — they watch any site. The Webflow-specific question is which tools fit your stack and can publish into it, not just report on it.
At a glance
## GEO tools for Webflow, compared
Grouped by the job you are hiring the tool to do. Pricing is indicative — check each vendor for current plans.
| Tool | What it does | Webflow fit | Pricing |
| --- | --- | --- | --- |
| HarperFlow | Writes & publishes cited GEO articles to your CMS | Native — publishes into Webflow | From $62/mo |
| Profound | Enterprise AI-visibility analytics & monitoring | Platform-agnostic tracker | From \~$99/mo |
| Otterly.ai | AI-search brand-mention monitoring | Platform-agnostic tracker | From \~$29/mo |
| Peec AI | AI visibility tracking across engines | Platform-agnostic tracker | Paid, tiered |
| AthenaHQ | AI brand-visibility monitoring | Platform-agnostic tracker | Paid, tiered |
| SE Ranking | SEO suite + AI-visibility tracker + audit | Works with any site | Paid, tiered |
| Semrush | SEO suite + AI/GEO toolkit | Works with any site | Paid, tiered |
Reviewed
## The tools, reviewed
What each one is genuinely good at — and where it does not fit.
### HarperFlow — create & publish cited content
The only tool here that writes and publishes finished, cited GEO articles straight into your Webflow CMS. It researches with citations, structures for extraction, and publishes on a schedule with a human veto. Best when your bottleneck is producing content built to be cited, not just measuring it. It is a content engine, not a tracker — pair it with one. [See how it works](https://www.harperflow.io/features).
### Profound — enterprise AI-visibility analytics
The enterprise standard for measuring how often AI engines mention and cite you, with competitor benchmarking. Platform-agnostic and strong for larger teams; its depth and pricing are aimed at that market.
### Otterly.ai — affordable AI-search monitoring
Tracks brand mentions and citations across ChatGPT, Perplexity, and Google AI at an accessible price. A good entry point if you mainly want to watch your visibility.
### Peec AI — AI visibility tracking
Monitors how your brand appears across engines with clear dashboards and competitor views. A solid tracker in a crowded field; evaluate on coverage and price.
### AthenaHQ — AI brand-visibility monitoring
Focuses on measuring and improving how AI assistants represent your brand. Another capable tracker — weigh it against Profound and Peec on the engines it covers.
### SE Ranking — SEO plus AI visibility
A full SEO suite that added an AI-visibility tracker and site audit. Handy if you want traditional SEO and AI monitoring in one place; works with any site, including Webflow.
### Semrush — the heavyweight suite
A broad SEO platform with an AI/GEO toolkit. Powerful but heavier and pricier than single-purpose trackers; overkill if you only need AI-visibility monitoring.
How to choose
## How to choose the right one
### Start with the job, not the brand
Decide whether your gap is measuring visibility, auditing pages, or creating cited content. Most teams need a tracker plus a way to produce content.
### Measuring is not earning
A tracker tells you AI ignores you; it does not fix it. Only \~31% of cited pages carry original data — creating that content is what moves the needle. [More on trackers vs earners](https://www.harperflow.io/ai-search-optimization-tools).
### On Webflow, favour fit
Trackers are platform-agnostic, so choose on coverage and price. For creating content, pick a tool that publishes into Webflow so you are not copy-pasting. [See the Webflow playbook](https://www.harperflow.io/how-to-get-your-webflow-site-cited-by-ai).
## Most GEO tools *measure*. Few help you *earn* the citation.
Track your visibility, yes — but budget for creating the content that changes it.
FAQ
## GEO tools for Webflow: common questions
What is a GEO tool?
A tool that helps your brand appear in AI answers. Some track your visibility across engines, some audit your pages, and some create content built to be cited. GEO stands for generative engine optimization.
Which GEO tools work with Webflow?
Trackers like Profound, Otterly, Peec AI, and AthenaHQ are platform-agnostic, so they work with any site including Webflow. For creating and publishing content into Webflow, HarperFlow is built for the stack.
Do I need a GEO tool if I already do SEO?
SEO and GEO overlap but are not the same. If you want to appear in ChatGPT, Perplexity, and Google's AI answers, a tracker shows where you stand and a content tool helps you earn citations.
What is the difference between a GEO tracker and a content engine?
A tracker measures whether AI mentions you; a content engine creates and publishes the pages that earn those mentions. You usually need both.
Are these GEO tools free?
Most offer a free trial or a limited free tier, with paid plans for real use. Prices range widely, from around $29/mo for entry trackers to enterprise pricing for platforms like Profound. HarperFlow starts at $62/mo.
Is HarperFlow a GEO tracker?
No. HarperFlow is a content engine — it writes and publishes cited GEO articles to your Webflow site. Pair it with a tracker if you also want to measure your AI visibility.
Compare the tools
## See how HarperFlow compares, tool by tool
[Surfer SEO alternative](https://www.harperflow.io/surferseo-alternative)· [Clearscope alternative](https://www.harperflow.io/clearscope-alternative)· [Frase alternative](https://www.harperflow.io/frase-alternative)· [MarketMuse alternative](https://www.harperflow.io/marketmuse-alternative)· [Copy.ai alternative](https://www.harperflow.io/copyai-alternative)· [Writesonic alternative](https://www.harperflow.io/writesonic-alternative)
Related guides: [free GEO audit](https://www.harperflow.io/geo-website-audit)· [LLM SEO](https://www.harperflow.io/llm-seo)
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
## See where your Webflow site stands
Run a free GEO audit, then decide which tools you actually need.
[Run a free GEO audit→](https://www.harperflow.io/geo-website-audit) [Start free](https://app.harperflow.io/auth/register?credits=10)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/seo-automation
# SEO Automation That Publishes,Not Just *Audits*.
Most SEO automation hands you a longer to-do list. It crawls your site, flags what's broken, tracks your rankings — and then leaves the actual work, writing the pages and getting them live on a schedule, to you. HarperFlow closes that loop. It runs [SEO automation on autopilot](https://www.harperflow.io/features): discover topics, draft the article, schedule it, email you a veto digest, then publish straight to your Webflow, WordPress, or Shopify blog. The audit is the easy part. Publishing good content on time, every week, is the part that never gets done.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [Start free today →](https://app.harperflow.io/auth/register)
Keyword and SERP data measured **27 July 2026** via Semrush (US).
Can it be automated?
## What can and can't be automated
*Can SEO be automated?* Parts of it, cleanly — and parts of it never should be.
### The mechanical work automates well
Keyword and topic discovery, first-draft writing, internal linking, image selection, scheduling, and the publish step itself are all repeatable enough for software to own. This is where search engine optimization automation earns its keep: it removes the steps that stall simply because a person has to remember to do them.
### The judgment work stays with you
Strategy, positioning, brand voice, and the final go/no-go do not automate, and any honest auto SEO tool will say so. HarperFlow's answer is a review window rather than a black box: every run creates scheduled drafts, you receive a digest before anything publishes, and you can veto or reschedule any piece inside that window. Nothing goes live that you didn't have the chance to stop.
The honest answer
So the honest answer to "can SEO be automated" is this: the production line can be, the editorial calls can't, and a good system keeps you in control of the second while it quietly runs the first.
Compared
## Automated SEO software, compared
Most automated SEO software is built to diagnose. It audits technical issues, tracks keyword positions, and suggests fixes — genuinely useful work, and where these tools are strongest. What most SEO automation tools don't do is finish the job: turn a recommendation into a published, scheduled article sitting live on your CMS.
Here is the honest split between the category and how HarperFlow works — we keep comparisons like this [sourced, dated, and fair](https://www.harperflow.io/compare), including where other tools genuinely win:
| Capability | Most automated SEO software | HarperFlow |
| --- | --- | --- |
| Technical audits and issue flagging | YesCore strength | PartialNot the core product |
| Keyword and topic discovery | Yes | Yes |
| Drafts full articles | Sometimes | Yes |
| Publishes directly to your CMS | Rare | YesWebflow, WordPress & Shopify |
| Schedules a publishing cadence | No | Yes |
| Human veto before anything goes live | n/a | YesVeto window |
| Grounds claims in named sources | No | Yes |
If you need a technical crawler, the category tools win — HarperFlow won't replace them and isn't trying to. If your bottleneck is that nothing gets written and published, that's the gap HarperFlow was built for. Comparing publishing autopilots specifically? See our [Soro alternative](https://www.harperflow.io/soro-alternative) for an honest, dated head-to-head.
The gap
## The publishing gap
Audits are abundant; published pages are scarce. Almost every team has a backlog of recommendations no one has had time to act on. Automation that stops at the recommendation just makes that backlog longer, faster — you end up with a more precise list of work that still isn't getting done.
The publishing gap is also where automated SEO increasingly decides whether AI answers cite you at all. [HarperFlow's own research](https://www.harperflow.io/state-of-ai-citations) found that how a page is structured predicts whether AI engines cite it, at a statistically significant level ([p=0.004](https://www.harperflow.io/can-your-site-get-cited-by-ai)). Content that is well-structured, dated, and genuinely live is what gets pulled into answers; a draft stuck in a backlog gets pulled into nothing, [no matter how strong the recommendation behind it](https://www.harperflow.io/can-your-site-get-cited-by-ai). Closing the distance between "recommended" and "published" is the entire point.
That is the part HarperFlow automates end to end: it writes structured, source-grounded articles and publishes them on a schedule to your Webflow, WordPress, or Shopify blog, so the calendar fills itself while you keep the veto.
## Auto SEO that only audits leaves the hardest step for you; automation that *publishes* actually gives the time back.
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions
## SEO automation, answered
Can SEO be automated?
The repetitive production work — discovery, drafting, internal linking, scheduling, and publishing — automates well. Strategy and final approval should not be fully hands-off. HarperFlow automates the production line and keeps you in the loop with a veto window before anything publishes.
What does automated SEO software actually do?
Most automated SEO software audits your site, tracks rankings, and recommends fixes. Fewer tools write the content, and very few publish it for you. HarperFlow is built for that last mile — drafting and publishing finished articles directly to your CMS.
Do SEO automation tools publish content for you?
Most don't; they stop at recommendations. HarperFlow does — it schedules and publishes finished articles straight to your Webflow, WordPress, or Shopify blog, after you have had the chance to veto anything in the digest.
## Automate the part that never gets done
Before you automate anything, find out whether AI answers already mention your site. When you are ready to publish on autopilot, HarperFlow's Flex plan starts at [$62/mo for Webflow and WordPress for 5 articles, billed monthly](https://www.harperflow.io/pricing) — about one published article every six days, hands-free. No legacy tiers, no per-seat add-ons, and no long-term contract; the plan renews at the amount you choose until you change it.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [See Flex pricing](https://www.harperflow.io/pricing)
Keep exploring
## Related reading
[Product
How HarperFlow works
Discover, draft, schedule, veto, publish — the autopilot loop end to end. Read →](https://www.harperflow.io/features) [Overview
All comparisons
Every AI content tool, compared side by side — sourced, dated, and fair. Read →](https://www.harperflow.io/compare) [Research
The state of AI citations
What actually predicts whether AI answers cite your pages. Read →](https://www.harperflow.io/state-of-ai-citations)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/webflow-seo
# Webflow SEO: The Complete Setup(and What Webflow Won't Do For You)
Webflow gives you one of the cleanest technical foundations in web publishing — semantic markup, fast hosting, and editable metadata on every page. But a clean foundation is not a ranking, and it is certainly not a citation in an AI answer. This is the honest version of Webflow SEO: what the platform does for you out of the box, the checklist that closes the gap, exactly where it stops, and how to keep new content shipping after launch.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [Start free today →](https://app.harperflow.io/auth/register)
Keyword and SERP data measured **27 July 2026** via Semrush (US).
The honest answer
## Is Webflow good for SEO?
The short answer
Yes — for the technical layer, Webflow is genuinely good, and better than most no-code builders. You get clean semantic HTML, global CDN hosting, automatic SSL, editable title tags and meta descriptions, a generated sitemap and robots.txt, canonical tags, 301 redirects, and image alt fields. Google has no trouble crawling or rendering a Webflow site.
### Where it stops
Where Webflow is not good — and where any honest guide should say so — is the part after setup. Webflow will not decide what to write, write it, publish it on a cadence, or earn the links and citations that actually move rankings. It hands you a fast, well-formed empty room. What goes in it, and how often, is on you.
### Two questions, not one
That distinction matters more in 2026 than it did five years ago, because "good for SEO" now splits into two questions: can Google index it (yes), and will an AI answer cite it (a separate problem the platform does not touch).
Out of the box
## What Webflow handles for you automatically
These are the technical SEO items you would otherwise wire up by hand or with plugins. On Webflow they ship with the platform.
- **Clean, semantic HTML.** Webflow outputs standards-compliant markup with real heading tags, so crawlers and AI engines parse your structure without a rendering fight.
- **Fast global hosting on a CDN.** Pages are served from a global CDN with automatic SSL, which protects Core Web Vitals without a caching plugin.
- **Editable titles and meta descriptions.** Every static page and CMS template exposes its own title tag and meta description, including fields you can bind to CMS data.
- **Automatic XML sitemap.** Webflow generates and maintains your sitemap as you add pages and CMS items; you just submit it once in Search Console.
- **Canonical tags and robots control.** Set canonical URLs on templates and edit robots.txt to steer crawling and avoid duplicate-URL dilution.
- **301 redirects.** Change a slug and map the old URL to the new one in Site Settings, preserving link equity.
- **Open Graph and image alt fields.** Social preview tags and per-image alt text are first-class fields, not afterthoughts.
- **Responsive, compressed images.** Webflow serves responsive image variants and modern formats automatically, trimming page weight on mobile.
In other words, the entire technical-hygiene layer of an SEO checklist is handled — genuinely more than most no-code builders give you. It is also exactly where Webflow's help stops.
The gap
## Where Webflow's SEO stops
A technically perfect site can still rank nowhere. These are the limits the platform will not close for you.
- **It doesn't choose topics.** Webflow has no idea which keywords or questions your buyers search. Topic strategy is entirely on you.
- **It doesn't write or structure content.** The CMS is an empty vessel; the words, the sources, and the on-page structure that earns citations are yours to produce.
- **Structured data beyond the basics is manual.** Open Graph is built in, but Article, FAQ, HowTo, and Product schema must be hand-added as JSON-LD in custom code, per template.
- **Internationalization is limited.** Native hreflang and multi-locale SEO are only partially handled and often need Localization or workarounds for larger sites.
- **CMS item and collection limits.** Plans cap CMS items and collections, which matters if you plan a large programmatic-content library.
- **It never publishes on a schedule.** The biggest ranking factor over time — shipping fresh, structured content consistently — is a habit no platform performs for you.
Every item on this list is a job, not a setting. You can do them by hand, hire them out, or automate the publishing layer with software — which is the honest case for a tool like HarperFlow.
Webflow vs WordPress
## Webflow SEO vs WordPress: which ranks better?
The most common question from teams weighing a rebuild. The honest answer: the platform sets your floor, not your ceiling.
### The honest comparison
Both platforms can rank at the top of Google, and both can sit invisible. Webflow gives you a cleaner technical baseline out of the box — semantic markup, fast global hosting, automatic SSL, sitemaps, canonical tags, and redirects — with none of the plugin sprawl that slows a typical WordPress install. WordPress gives you a deeper plugin ecosystem such as Yoast or RankMath and more raw flexibility, at the cost of maintenance, security patching, and performance tuning you own forever.
### Where Webflow wins
If you want a fast, stable site where the SEO fundamentals are handled without five plugins and a caching config, Webflow is the lower-effort baseline — and its clean HTML is exactly what AI engines parse when they decide what to quote. You spend your time on content, not infrastructure.
### Where WordPress still wins
Very large content libraries, complex multi-author workflows, or a hard dependency on a specific WordPress plugin can still tip the decision. But for most marketing sites the deciding factor isn't the CMS — it's whether anyone is publishing structured, cited content on a schedule. That is the gap HarperFlow closes on Webflow: it researches, writes, cites, and publishes straight to your CMS.
Free checklist
## The Webflow SEO checklist
Use this as a free SEO checklist you can run against any Webflow project before and after launch. It is deliberately short and ordered by impact.
1. Set a unique title tag and meta description on every page and CMS template — not just the homepage.
2. Define one clear H1 per page, and use H2s for real section structure rather than styling.
3. Fill the alt text field on every image, describing the actual image.
4. Turn on the auto-sitemap, submit it in Google Search Console, and confirm robots.txt is not blocking anything live.
5. Set canonical tags on CMS templates to avoid duplicate-URL dilution.
6. Add Open Graph title, description, and image for clean link previews.
7. Compress images and lean on Webflow's responsive image handling — Core Web Vitals count.
8. Add structured data (Article, FAQ, Breadcrumb) where it fits; Webflow renders it fine inside an embed.
9. Wire up 301 redirects for any URL you change.
10. Publish new, well-structured content on a schedule — the item most checklists skip, and the one that compounds.
Items 1 through 9 are one-time hygiene. Item 10 is the work, and it never ends.
### How to knock out the high-impact items
Titles and metas: write them per page and per CMS template, leading with the term you want to rank for. Canonical tags: set them on every collection template so filtered and paginated URLs don't split your ranking signal. Sitemap: toggle Webflow's auto-sitemap on and submit it once in Search Console. Structured data: paste Article, FAQ, and Breadcrumb JSON-LD into each template's custom code. And the one that compounds — publishing cadence: decide how many pieces you will ship each month before launch, because after launch it is the only item on this list that keeps paying off.
## Items 1–9 are one-time hygiene. Item 10 — *publishing on a schedule* — is the work.
And it's the one item you can hand to software instead of a retainer.
Handled vs left to you
## What Webflow handles vs what it doesn't
The left column is why people choose Webflow. The right column is why a Webflow site can sit technically perfect and still rank nowhere.
| What Webflow handles for you | What Webflow leaves to you |
| --- | --- |
| YesClean semantic HTML and fast CDN hosting | NoPublishing new content on a schedule |
| YesEditable titles, metas, and OG tags | NoWriting those tags for every page and CMS item |
| YesAuto-generated sitemap and robots.txt | NoChoosing the topics worth targeting |
| YesCanonical tags, 301s, and alt-text fields | NoActually filling the alt text and canonicals in |
| YesSSL, clean URLs, responsive rendering | NoEarning the links and AI citations that rank you |
| YesAI Assistant SEO suggestions | You judgeWhether a suggestion fits your niche |
Hiring help
## Hiring help: agencies and services
If the right column is not your job to do, you hire it out. There are three tiers of help, and the labels matter.
01
### Ongoing and retained
A **Webflow SEO company** or **Webflow SEO agency** typically owns strategy, content, and ongoing publishing — a retained relationship.
02
### Scoped and one-off
**Webflow SEO services** are usually scoped, one-off engagements: a technical audit, a migration, or a single checklist pass.
03
### Hire on the axis you care about
We keep a public, sourced directory of 43 vetted [Webflow agencies](https://www.harperflow.io/webflow-agencies), and — because "ranks in Google" and "gets cited by AI" are now different skills — we separately measure which of them actually surface in AI answers in the [Webflow Agency AI Visibility Index](https://www.harperflow.io/webflow-agency-ai-visibility-index). Use both: hire the shop that is strong on the axis you care about, and be honest with yourself about which axis that is.
Automate the cadence
## Automating it
The one checklist item that compounds — publishing structured content on a schedule — is also the one you can hand to software instead of a retainer.
### Webflow-first autopilot
That is what HarperFlow does, and it is Webflow-first: it runs on autopilot to discover topics, draft GEO-optimized articles, schedule them, email you a veto digest, and publish the un-vetoed drafts straight to your Webflow CMS.
### Built to be cited
The AI-citation angle is deliberate. HarperFlow's own analysis of AI answers found that how a page is *structured* predicts whether it gets cited — a real, measured effect (p=0.004), documented in [Can your site get cited by AI](https://www.harperflow.io/can-your-site-get-cited-by-ai). Every article HarperFlow publishes is built to that structural standard, and each one carries a human veto window before it goes live, so automation never means losing control. See how the pipeline fits together in [Features](https://www.harperflow.io/features), or read the setup path in the [docs](https://www.harperflow.io/docs).
### Monthly, usage-based pricing
Pricing is monthly and usage-based: the Flex plan starts at **$62/mo for 5 articles** ($12.40 each) and scales up to 30, with no legacy tier lock-in. Full detail on [pricing](https://www.harperflow.io/pricing).
Keep reading
## Related Webflow guides
Two companion guides go deeper: [Webflow AI — how to use AI for your content](https://www.harperflow.io/webflow-ai) without shipping slop, and [The Webflow Blog Guide](https://www.harperflow.io/webflow-blog) on setting up and running a blog that gets cited. Together with this page, they cover SEO, AI, and publishing on Webflow end to end.
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions
## Frequently asked questions
Is Webflow good for SEO?
Yes, for the technical foundation — semantic HTML, fast hosting, editable metadata, sitemaps, and redirects all ship in the box. It does not choose topics, write content, publish on a schedule, or earn citations; those remain your work, or your agency's.
What should be on a Webflow SEO checklist?
Unique titles and metas per page and CMS template, one clear H1, filled alt text, an active sitemap in Search Console, canonical tags, Open Graph tags, compressed images, structured data, 301 redirects, and — the item most lists omit — a real publishing cadence.
Do I need a Webflow SEO agency, or can I do it myself?
The one-time technical checklist is very DIY-able on Webflow. Ongoing topic strategy, writing, and publishing are where a Webflow SEO agency or scoped services earn their fee — unless you automate that cadence with software instead.
How much do Webflow SEO services cost?
Retained Webflow SEO companies run into the thousands per month; scoped services are billed per project. Automating the publishing layer costs far less — HarperFlow's Flex plan starts at $62/mo for 5 articles.
Does Webflow create a sitemap automatically?
Yes. Webflow generates and keeps an XML sitemap updated for you, and you can toggle it under Site Settings then SEO, or hand-author one if you need custom control. Submit it once in Google Search Console and Webflow keeps it current as you add pages and CMS items.
How do I add schema markup in Webflow?
Webflow doesn't generate structured data on its own beyond basic Open Graph tags. You add JSON-LD by pasting it into a page's custom code (Page Settings, then head) or a site-wide embed — Article, FAQPage, and Breadcrumb are the highest-value types. If you publish at volume, generate the schema per template so every new post ships with it, which is exactly what HarperFlow automates.
Is Webflow or WordPress better for SEO?
Neither wins on ranking by default — both can rank, and both can flop. Webflow gives you a cleaner technical baseline and fast hosting with no plugin sprawl; WordPress gives you Yoast or RankMath and a bigger ecosystem, but more maintenance and performance risk. On Webflow the ceiling isn't the platform, it's whether you publish structured, cited content on a schedule.
Does Webflow SEO work without any apps or plugins?
Yes. Unlike WordPress, Webflow's SEO essentials — titles, metas, sitemaps, canonical tags, redirects, alt text, and clean markup — are built into the platform, so you don't need a plugin to cover the fundamentals. What no plugin or platform does for you is the ongoing work: choosing topics, writing, and publishing content that earns rankings and AI citations.
## Run a free GEO audit
Before you hire anyone or touch a line of markup, see where your Webflow site actually stands in AI answers — no login, sourced and dated like everything else we publish.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [See how HarperFlow works](https://www.harperflow.io/features)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/webflow-ai
# Webflow AI: What's Real, and How to Use It Well
AI can draft a Webflow page in seconds — and fill the web with forgettable content just as fast. This is the honest version of Webflow AI: what Webflow's own AI actually does, the third-party tools that write and publish for you, and how to use AI on your Webflow site so the result gets read and cited instead of ignored.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [Start free today →](https://app.harperflow.io/auth/register)
Researched and updated **August 2026**. Every claim checked — no AI slop.
The honest answer
## What is Webflow AI?
The short answer
"Webflow AI" means two different things. Webflow ships its own AI features — an AI Assistant that helps build and edit pages, AI-assisted site generation, and AI help inside the Designer. Separately, and more important for growth, there is the wider category of AI tools that research, write, and publish the content that lives on your Webflow site. The first builds the container; the second fills it — and that is where most of the ranking and AI-citation upside actually sits.
### Webflow's native AI
Inside Webflow, AI mostly accelerates building: generating layouts, editing copy in place, and answering how-to questions in the Designer. It is a genuine time-saver for construction. What it does not do is decide what to publish, ground claims in sources, or keep a blog shipping every week.
### AI for your Webflow content
The higher-leverage use of AI on Webflow is content: turning a topic into a researched, structured, cited article and getting it into your CMS. Done well, this is how a Webflow site earns search traffic and shows up in AI answers. Done badly, it is how the internet filled with slop.
The capabilities
## What AI can actually do for a Webflow site today
Beyond building pages, here is where AI genuinely helps a Webflow site grow — and where a human still has to steer.
- **Draft and edit page copy.** Webflow's AI Assistant and general writing tools speed up headlines, product copy, and edits directly in the Designer.
- **Research and outline articles.** AI can gather sources, cluster subtopics, and propose a structure far faster than a blank page — the raw material of a real post.
- **Write full, structured drafts.** Modern models produce complete articles with headings, FAQs, and tables — the formats AI engines prefer to quote.
- **Generate schema and metadata.** AI can output JSON-LD, titles, and meta descriptions per template so every CMS item ships search-ready.
- **Publish straight to the CMS.** With the Webflow API, a pipeline can push finished, formatted articles into your Collections on a schedule — no copy-paste.
- **Track AI-search visibility.** Tools now measure whether ChatGPT, Perplexity, and Gemini cite you, closing the loop on what to write next.
Every one of these is real. The catch is that AI will do all of them badly by default — unless someone insists on sources, structure, and a veto.
The slop problem
## Where AI on Webflow goes wrong
The reason "AI content" has a bad name is that most of it is generated, not researched — and search engines have noticed.
- **Unsourced claims.** A model that isn't grounded in real citations states things confidently and wrongly, which is exactly what AI search engines are learning to filter out.
- **Sameness at scale.** Ten AI articles on the same prompt read like one article ten times. Volume without a point of view earns nothing.
- **No structure that gets cited.** Walls of text don't get quoted. FAQ blocks, tables, and clear headings do — and most AI drafts skip them.
- **No human veto.** Publishing whatever the model produced, unread, is how factual errors and off-brand claims reach your live site.
We measure this failure mode in depth in our [Dead Internet Report](https://www.harperflow.io/dead-internet) — the collapse of properties AI has commoditized. The lesson for a Webflow site is simple: AI is only an asset when it is researched, cited, and checked.
The three approaches
## DIY prompting vs Webflow's AI vs a publishing pipeline
Three ways to put AI to work on a Webflow blog, and what each is actually good at.
### DIY prompting
Cheap and flexible: you prompt a chatbot, paste the result into Webflow, and edit. Fine for a one-off, but it doesn't scale — every article is manual, sources are your job, and quality drifts with your energy.
### Webflow's native AI
Best for building and editing the site itself. It speeds up construction inside the Designer, but it isn't a content engine — it won't research a topic, cite sources, or publish on a cadence.
### A publishing pipeline
Purpose-built for content: it researches with citations, writes in cited and structured formats, runs an independent audit, keeps a human veto, and publishes straight to your Webflow CMS on a schedule. This is the lane HarperFlow is built for.
The playbook
## How to use AI for Webflow content without shipping slop
A short, honest checklist for getting the upside of AI without the reputational cost.
1. Start from a real keyword or buyer question, not a generic prompt.
2. Make the model cite sources for every non-obvious claim, and check them.
3. Write in citable formats — clear H2s, an FAQ block, and at least one table.
4. Add Article and FAQ schema so engines can parse and quote the page.
5. Keep a human veto: read it before it publishes, and kill anything off or wrong.
6. Publish on a schedule, then measure whether AI engines actually cite you.
Do these six by hand, hire them out, or automate the whole loop — that last option is the honest case for using [HarperFlow](https://www.harperflow.io/features) on Webflow.
## The web doesn't need more AI content. It needs *content worth citing*.
That's the only kind AI search rewards — and the only kind worth publishing.
Automate the cadence
## Automating it
The one checklist item that compounds — publishing structured content on a schedule — is also the one you can hand to software instead of a retainer.
### Webflow-first autopilot
That is what HarperFlow does, and it is Webflow-first: it runs on autopilot to discover topics, draft GEO-optimized articles, schedule them, email you a veto digest, and publish the un-vetoed drafts straight to your Webflow CMS.
### Built to be cited
The AI-citation angle is deliberate. HarperFlow's own analysis of AI answers found that how a page is *structured* predicts whether it gets cited — a real, measured effect (p=0.004), documented in [Can your site get cited by AI](https://www.harperflow.io/can-your-site-get-cited-by-ai). Every article HarperFlow publishes is built to that structural standard, and each one carries a human veto window before it goes live, so automation never means losing control. See how the pipeline fits together in [Features](https://www.harperflow.io/features), or read the setup path in the [docs](https://www.harperflow.io/docs).
### Monthly, usage-based pricing
Pricing is monthly and usage-based: the Flex plan starts at **$62/mo for 5 articles** ($12.40 each) and scales up to 30, with no legacy tier lock-in. Full detail on [pricing](https://www.harperflow.io/pricing).
Keep reading
## Related Webflow guides
Two companion guides go deeper: [Webflow SEO — the complete setup](https://www.harperflow.io/webflow-seo) and what the platform won't do for you, and [The Webflow Blog Guide](https://www.harperflow.io/webflow-blog) on building a blog that gets read and cited. Together they cover SEO, AI, and publishing on Webflow end to end.
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions
## Frequently asked questions
Does Webflow have AI?
Yes. Webflow ships AI features inside the platform, including an AI Assistant for building and editing pages and AI-assisted site generation. They speed up construction; for content that ranks and gets cited, most teams pair Webflow with a dedicated AI writing or publishing tool.
What is Webflow's AI Assistant?
It's Webflow's built-in helper in the Designer that can generate and edit layouts and copy and answer how-to questions as you build. It is aimed at building the site faster — not at researching topics, citing sources, or publishing a blog on a schedule.
Can I use AI to write my Webflow blog?
Yes, and many sites do. The key is grounding the draft in real sources, structuring it with headings, FAQs, and tables, and reading it before it publishes. AI that just generates unsourced text tends to hurt more than it helps.
Is AI content bad for SEO?
AI content isn't penalized for being AI-written; thin, unsourced, duplicative content is — whoever wrote it. Google and AI engines reward useful, original, well-structured pages. Used with citations and a human veto, AI is a legitimate way to produce them.
What are the best AI tools for Webflow content?
It depends on the job. Webflow's native AI is best for building the site; general writers like ChatGPT or Jasper help draft; and publishing pipelines like HarperFlow research, cite, audit, and push finished articles straight to your Webflow CMS.
How do I use AI without publishing "AI slop"?
Start from a real question, force citations, write in citable formats, keep a human veto, and publish on a cadence you can sustain. Slop is what happens when you skip those steps and publish raw model output at volume.
## Run a free GEO audit
Before you hire anyone or touch a line of markup, see where your Webflow site actually stands in AI answers — no login, sourced and dated like everything else we publish.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [See how HarperFlow works](https://www.harperflow.io/features)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/webflow-blog
# The Webflow Blog Guide: Build It, Run It, Get It Cited
Webflow's CMS is one of the best places to run a blog — clean, fast, and fully designable. But a blog is only an asset if people, and AI engines, actually read and cite it. This is the practical guide: how to set up a Webflow blog, structure it to rank and get cited, and keep it shipping after launch.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [Start free today →](https://app.harperflow.io/auth/register)
Researched and updated **August 2026**. Built on Webflow CMS.
The honest answer
## Can you blog on Webflow?
The short answer
Yes — and it's one of Webflow's strongest use cases. You build a blog with a CMS Collection: define the fields once (title, slug, body, author, category, image, meta), design a single template, and every new post inherits it. You get clean URLs, RSS, categories, and full design control, with none of the plugin maintenance a WordPress blog demands. What Webflow won't do is fill it — the writing and the publishing cadence are still yours.
### Why the CMS matters
A Webflow blog isn't a pile of static pages; it's structured data. Because every post shares one Collection schema, you can bind SEO fields, schema markup, and Open Graph tags once and have them apply to every article — which is exactly what makes a blog scalable and citable.
### What a good Webflow blog needs
Three things beyond the design: a clear structure engines can parse, a steady flow of genuinely useful posts, and internal links that tie the blog to the rest of your site. Miss the last two and even a beautiful Webflow blog sits unread.
Setup
## How to set up a blog on Webflow
The full path from empty site to a live, structured blog — the one-time work you do before you ever write a post.
1. Create a CMS Collection called Blog Posts and add fields: title, slug, summary, body (rich text), author (reference), category (reference), main image, and meta title and description.
2. Design one Collection Page template — the layout every post will use — with the fields bound in place.
3. Add a Collection List on a blog index page to show recent posts, and enable pagination.
4. Create a Categories Collection and reference it from Blog Posts so readers and engines can browse by topic.
5. Bind the meta title, meta description, and Open Graph fields on the template so every post ships search-ready.
6. Turn on Webflow's RSS feed for the Collection and link it in your footer.
7. Set a canonical tag on the template and confirm the auto-sitemap includes your posts.
8. Publish, submit the sitemap in Google Search Console, and write your first post.
Done once, this structure carries every future article. From here, the job stops being setup and becomes publishing.
Get read, get cited
## How to make a Webflow blog rank and get cited by AI
A live blog is table stakes. These are the choices that decide whether it earns search traffic and shows up in AI answers.
- **Answer a real question per post.** Target one keyword or buyer question, and answer it directly in the first paragraph — the sentence engines quote.
- **Structure for citation.** Use clear H2s, an FAQ block, and at least one table. Cited pages are structured pages.
- **Ship schema per template.** Bind Article, FAQ, and Breadcrumb JSON-LD to the Collection template so every post is machine-readable.
- **Cite your sources.** Link to primary data. Grounded, sourced posts are what AI engines increasingly prefer to reference.
- **Interlink deliberately.** Link new posts to related articles and to your product and comparison pages so authority flows where it converts.
- **Keep it fresh.** Update cornerstone posts and bump the modified date; freshness is a ranking and citation signal Webflow won't manage for you.
None of these are Webflow settings. They are editorial decisions — the reason two identical Webflow blogs can perform completely differently.
The comparison
## Webflow blog vs WordPress and Ghost
If you're choosing a platform to blog on, here's the honest trade-off.
### Webflow
Best design control and a clean, fast, plugin-free foundation with SEO essentials built in. The trade-off is CMS item limits on lower plans and schema you add by hand. Ideal when the blog is part of a designed marketing site.
### WordPress
The biggest ecosystem and plugins like Yoast for everything — at the cost of maintenance, security, and performance work you own forever. Ideal for very large, plugin-dependent content operations.
### Ghost
A focused, fast writing-and-newsletter platform. Lovely for a pure publication, but weaker when the blog needs to live inside a fully designed product site. Webflow wins when design and marketing matter as much as the words.
The real work
## The hard part isn't setup — it's publishing every week
Most Webflow blogs don't fail on design. They fail because the second post never gets written.
### Cadence is the whole game
Search and AI visibility compound with consistent, structured publishing. One great post a quarter loses to a steady one a week. That cadence is a habit no template enforces — it's where blogs quietly die.
### Where HarperFlow fits
HarperFlow runs the cadence for you on Webflow: it researches a topic, writes a cited and structured article, runs an independent audit, keeps a human veto, and publishes straight into your CMS Collection. You get a blog that keeps shipping — [see how it works](https://www.harperflow.io/features), or [use AI on your blog without the slop](https://www.harperflow.io/webflow-ai).
## A blog isn't a launch. It's a *habit*.
The sites that win are the ones still publishing in month six.
Automate the cadence
## Automating it
The one checklist item that compounds — publishing structured content on a schedule — is also the one you can hand to software instead of a retainer.
### Webflow-first autopilot
That is what HarperFlow does, and it is Webflow-first: it runs on autopilot to discover topics, draft GEO-optimized articles, schedule them, email you a veto digest, and publish the un-vetoed drafts straight to your Webflow CMS.
### Built to be cited
The AI-citation angle is deliberate. HarperFlow's own analysis of AI answers found that how a page is *structured* predicts whether it gets cited — a real, measured effect (p=0.004), documented in [Can your site get cited by AI](https://www.harperflow.io/can-your-site-get-cited-by-ai). Every article HarperFlow publishes is built to that structural standard, and each one carries a human veto window before it goes live, so automation never means losing control. See how the pipeline fits together in [Features](https://www.harperflow.io/features), or read the setup path in the [docs](https://www.harperflow.io/docs).
### Monthly, usage-based pricing
Pricing is monthly and usage-based: the Flex plan starts at **$62/mo for 5 articles** ($12.40 each) and scales up to 30, with no legacy tier lock-in. Full detail on [pricing](https://www.harperflow.io/pricing).
Keep reading
## Related Webflow guides
Two companion guides go deeper: [Webflow SEO — the complete setup](https://www.harperflow.io/webflow-seo) and where the platform stops, and [Webflow AI](https://www.harperflow.io/webflow-ai) on using AI for your content without shipping slop. Together with this page, they cover SEO, AI, and publishing on Webflow end to end.
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions
## Frequently asked questions
Can you blog on Webflow?
Yes. You build a blog with a CMS Collection: define your post fields once, design a single template, and every new article inherits the layout, SEO fields, and schema. You get clean URLs, categories, and RSS without any plugins.
Is Webflow good for blogging?
Very — especially when the blog is part of a designed marketing site. You get top-tier design control, fast hosting, and built-in SEO essentials. The trade-offs are CMS item limits on lower plans and schema you add by hand.
How do I set up a blog in Webflow?
Create a Blog Posts CMS Collection with your fields, design one Collection Page template, add a Collection List for the index, wire up categories, bind the SEO and Open Graph fields, enable RSS, and publish. After that, every post reuses the structure.
How much does a Webflow blog cost?
A blog needs a Webflow CMS site plan, which starts around $23 per month billed annually, plus your workspace plan. The bigger cost is content: writing and publishing consistently, which you can do in-house, hire out, or automate.
Can I migrate my WordPress blog to Webflow?
Yes. Export your WordPress posts to CSV, map the columns to a Webflow CMS Collection, and import. Set 301 redirects from the old URLs to the new ones so you keep your rankings. Larger blogs may use a migration tool or the Webflow API.
Does a Webflow blog have RSS?
Yes. Webflow generates an RSS feed for any CMS Collection. Enable it in the Collection settings and link it in your footer so readers and aggregators can subscribe.
How do I get my Webflow blog cited by AI?
Answer a clear question per post, structure it with headings, an FAQ, and a table, add Article and FAQ schema, cite your sources, and publish consistently. AI engines quote structured, sourced pages — which is exactly what HarperFlow is built to produce on Webflow.
## Run a free GEO audit
Before you hire anyone or touch a line of markup, see where your Webflow site actually stands in AI answers — no login, sourced and dated like everything else we publish.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [See how HarperFlow works](https://www.harperflow.io/features)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/webflow-cms
# Webflow CMS: How It Works, Where It Bites, When to Use It
Webflow's CMS is one of the best visual content systems you can build on — and, like everything, it has real limits no landing page mentions. This is the honest, hands-on guide: how Collections actually work, the caps and pricing that bite, the API we publish through every day, and when Webflow CMS is genuinely the wrong tool.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [Start free today →](https://app.harperflow.io/auth/register)
Researched and updated **August 2026**. We publish to Webflow CMS daily.
The honest answer
## What is Webflow CMS?
The short answer
Webflow CMS is the content layer of Webflow — a database of reusable content, called Collections, that you design once and populate many times. Instead of hand-building every blog post or product page, you define a Collection's fields, design one template, and Webflow generates a page for each item, with clean URLs, dynamic lists, and built-in SEO fields. It is a visual, no-code CMS aimed at marketing sites, not complex applications.
### Content, separated from design
The whole point of a CMS is to store content once and reuse it everywhere. In Webflow, an editor updates a Collection item and every page that references it updates too — no touching the design, no copy-paste across pages.
### Where it fits
It shines for blogs, portfolios, job boards, event listings, team pages, and product catalogs — anything that is a repeating pattern of structured content on a designed marketing site. It is not built to be the backend for a data-heavy web app.
How it works
## How Webflow CMS actually works
Five concepts do all the work. Once they click, the CMS is genuinely fast to build with.
- **Collections** are your content types — Blog Posts, Authors, Categories. Think of each as a database table.
- **Collection items** are the records: one blog post is one item, with its own values for every field.
- **Fields** define an item's shape — plain text, rich text, images, links, dates, switches, option sets, and reference fields, up to around 30 per Collection.
- **Reference & multi-reference fields** link Collections together — a post to an author, a post to several categories — so you can build related content without duplication.
- **Collection Page templates** are designed once; every item renders through that one template. **Collection Lists** then pull items anywhere on the site, with filtering and sorting, and any field can be **bound** to any element — text, image, link, or even a meta tag.
Items go live when you publish, or programmatically through the API. That is the entire model — and it is why a well-structured Webflow blog scales without extra work.
The part nobody mentions
## Where Webflow CMS bites you: the real limits
The feature pages skip these. Know them before you commit a large project.
- **Item caps by plan.** The CMS site plan caps CMS items (in the low thousands); the Business plan raises it, but very large libraries can still hit the ceiling.
- **Collection and field caps.** There is a limit on Collections per site and reference fields per Collection — fine for most sites, tight for complex content models.
- **Shallow relationships.** References are simple links, not a relational database. Modelling app-like data with deep joins is the wrong job for it.
- **Localization costs extra.** Multi-language is a paid Localization add-on, not native — a real cost for international content.
- **No native versioning or rollback** for individual CMS items, and no true repeating sub-structures within an item.
- **API rate limits.** The Data API is generous but throttled; bulk publishing has to be paced (we know — we do it daily).
None of these are dealbreakers for a marketing site or blog. They are exactly the things to check before you migrate ten thousand records into it.
Pricing, honestly
## What Webflow CMS actually costs
The CMS isn't a standalone product — it's unlocked by a site plan. Here's the real shape of it.
- **CMS site plan** (around $23/month billed annually) — unlocks the CMS with a few thousand items and more form submissions. The right tier for most blogs and marketing sites.
- **Business site plan** (around $39/month annually) — raises the item cap and bandwidth for larger, higher-traffic content sites.
- **Enterprise** — custom pricing for scale, security, and guarantees.
- **Workspace plan** — a separate cost if you build for clients or as a team, on top of the per-site plan.
- **Localization** — a paid add-on; only pay for it if you genuinely publish in multiple languages.
The honest summary: pick the site-plan tier that matches your item count and traffic, and ignore the add-ons you don't need. The bigger cost of any blog isn't the plan — it's producing content consistently.
The API
## The Webflow CMS API, from a team that uses it daily
This is where Webflow CMS stops being a place you hand-fill and becomes something that can fill itself.
- **It's a REST Data API with token auth.** You can create, read, update, and delete items, manage Collections and fields, and trigger publishes — all programmatically.
- **Staged vs live.** Items can be created as drafts and published later, or published live in a single call — which is how scheduled and reviewed publishing works.
- **It's rate-limited.** Generous, but throttled, so bulk jobs pace themselves rather than firing all at once.
- **It's the only real path to automation.** Scheduled publishing, syncing from another system, or pushing AI-generated articles into your blog — none of it happens without the API.
This is exactly the seam HarperFlow plugs into: it researches, writes, and cites an article, then publishes it straight into your Webflow CMS through this API — no copy-paste, on a schedule. [See how the pipeline works](https://www.harperflow.io/features).
Make it rank
## Structuring a Webflow CMS blog that ranks and gets cited
Webflow's "built-in SEO" is real but shallow — it gives you the fields; the structure is on you. Here's the framework.
- **Model the Collection for SEO.** Add separate meta title, meta description, and Open Graph fields, plus author and category reference fields.
- **Bind them on the template** so every post ships with a unique title, meta, and clean OG tags automatically.
- **Add schema per template.** Paste Article, FAQ, and Breadcrumb JSON-LD into the template's custom code so every item is machine-readable.
- **Structure the body to be cited.** A direct-answer intro, clear H2s, an FAQ block, and a table where it fits — the formats AI engines quote.
- **Interlink deliberately** and publish on a cadence you can sustain.
Get the Collection schema right once and every future post inherits it. We go deeper in [the Webflow blog guide](https://www.harperflow.io/webflow-blog) and [the Webflow SEO guide](https://www.harperflow.io/webflow-seo).
The comparison
## Webflow CMS vs WordPress and headless
Three honest comparisons — the modern landscape, not just WordPress.
### vs WordPress
Webflow gives design control and a clean, plugin-free foundation with SEO essentials built in. WordPress gives the biggest ecosystem and plugins like Yoast, at the cost of maintenance, security, and performance work. Webflow wins when design and a marketing site matter as much as the content.
### vs headless (Strapi, Sanity, Contentful)
Headless separates content from presentation entirely — powerful for multi-channel or app content, but you build and host the frontend yourself. Webflow bundles content, design, and hosting; you trade ultimate flexibility for speed and no-code.
### The honest call
For a designed marketing site or blog, Webflow CMS is usually the fastest path. For app-like data or genuine multi-channel publishing, headless wins. Match the tool to the job, not the hype.
Be honest with yourself
## When Webflow CMS is not the right fit
A guide that only sells the tool isn't a guide. Here's when to look elsewhere.
- **Massive catalogs** — tens of thousands of items that blow past the CMS item caps.
- **App-like relational data** with deep joins and complex queries — that's a database, not a CMS.
- **Heavy multi-language** at scale, where Localization costs add up fast.
- **Complex editorial workflows** — granular approvals, roles, and permissions beyond Webflow's basics.
- **True multi-channel** — the same content feeding a mobile app, a site, and more; that's a headless job.
If none of those describe you, Webflow CMS is very likely the right — and fastest — choice. And once your content is in it, the remaining problem is publishing consistently.
## A CMS doesn't write your content. It just makes *publishing it* painless.
The hard part — producing content worth publishing — is still yours. Or ours.
Automate the cadence
## Automating it
The one checklist item that compounds — publishing structured content on a schedule — is also the one you can hand to software instead of a retainer.
### Webflow-first autopilot
That is what HarperFlow does, and it is Webflow-first: it runs on autopilot to discover topics, draft GEO-optimized articles, schedule them, email you a veto digest, and publish the un-vetoed drafts straight to your Webflow CMS.
### Built to be cited
The AI-citation angle is deliberate. HarperFlow's own analysis of AI answers found that how a page is *structured* predicts whether it gets cited — a real, measured effect (p=0.004), documented in [Can your site get cited by AI](https://www.harperflow.io/can-your-site-get-cited-by-ai). Every article HarperFlow publishes is built to that structural standard, and each one carries a human veto window before it goes live, so automation never means losing control. See how the pipeline fits together in [Features](https://www.harperflow.io/features), or read the setup path in the [docs](https://www.harperflow.io/docs).
### Monthly, usage-based pricing
Pricing is monthly and usage-based: the Flex plan starts at **$62/mo for 5 articles** ($12.40 each) and scales up to 30, with no legacy tier lock-in. Full detail on [pricing](https://www.harperflow.io/pricing).
Keep reading
## Related Webflow guides
Put the CMS to work: [The Webflow Blog Guide](https://www.harperflow.io/webflow-blog) on building a blog that gets cited, [Webflow SEO](https://www.harperflow.io/webflow-seo) for the fundamentals, and [Webflow AI](https://www.harperflow.io/webflow-ai) on using AI for your content without the slop. New to AI search? Start with [generative engine optimization](https://www.harperflow.io/generative-engine-optimization).
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions
## Frequently asked questions
What is Webflow CMS?
Webflow CMS is Webflow's content layer: a database of reusable content called Collections. You design a template once and Webflow generates a page for every item, with clean URLs and built-in SEO fields. It's a visual, no-code CMS for marketing sites and blogs.
How much does Webflow CMS cost?
Webflow CMS is unlocked by a site plan — the CMS plan is around $23/month billed annually, and the Business plan (around $39/month) raises item and bandwidth limits. Building for clients also needs a separate Workspace plan, and Localization is a paid add-on.
What are Webflow CMS's limits?
The main ones: CMS item caps by plan (low thousands on CMS, more on Business), limits on Collections and reference fields, no deep relational modelling, no native item versioning, paid localization, and API rate limits. Fine for most blogs; check them before migrating a huge library.
Does Webflow CMS have an API?
Yes. Webflow's Data API is a REST API that lets you create, update, and publish CMS items programmatically — the same API automation tools, and HarperFlow, use to publish content straight into your Collections.
Can I export my content from Webflow CMS?
Yes. You can export CMS Collection data to CSV from the Designer, and read it via the API. Note that a CSV export is the content, not your design — plan a migration carefully if you're ever leaving Webflow.
How many CMS items can Webflow handle?
It depends on your plan. The CMS site plan supports items in the low thousands per site; the Business plan raises the cap substantially. For catalogs in the tens of thousands, confirm the current plan limits or consider a headless CMS.
Webflow CMS vs WordPress — which is better?
Neither wins by default. Webflow gives design control, fast hosting, and built-in SEO with no plugin maintenance; WordPress gives a bigger plugin ecosystem at the cost of upkeep. Webflow CMS usually wins when design and a marketing site matter as much as the content.
## Run a free GEO audit
Before you hire anyone or touch a line of markup, see where your Webflow site actually stands in AI answers — no login, sourced and dated like everything else we publish.
[Run a free GEO audit →](https://www.harperflow.io/geo-website-audit) [See how HarperFlow works](https://www.harperflow.io/features)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/geo-research
# Who AI recommends, and why.
We ask the engines real buyer questions and log exactly who they name and what they cite. Then we publish the raw data so you can check us. Everything below is measured — nothing here is estimated, illustrative, or rounded to look better.
## Every number is measured
No estimates, no "representative" figures. If we couldn't measure it, it isn't in the report.
## The dataset is public
The raw CSVs ship with every report. Recompute anything you doubt.
## We publish our failures
Hypotheses that didn't survive significance testing get named, not quietly dropped.
The 2026 AI Citations series
## Six reports, one method
Real buyer questions, run through ChatGPT, Perplexity and Gemini — first 39 questions on July 1, then 250 (including the identical 39) on July 16. Each report is a different true thing in the same measured data.
[No. 1
The 2026 State of AI Citations
56% Ask three engines the same buying question and all three share even one recommendation just 56% of the time. Their full lists overlap \~7%. "Rank on AI" is really three separate games.All 3 engines + sourcesRead →](https://www.harperflow.io/state-of-ai-citations) [No. 4
One Engine's Favourite, Everyone Else's Unknown
79% The most-recommended brand in the study is named only by ChatGPT — zero from Perplexity, zero from Gemini. And 79% of the 646 brands exist in exactly one engine's world.All 3 enginesRead →](https://www.harperflow.io/one-engines-favourite) [No. 2
In Every Room, Never on the Podium
87% / 0 Reddit is cited in 87% of Perplexity's answers — and was the top source for none of them. Meanwhile 86% of cited domains appeared exactly once. The tail is the market.Perplexity only · n=39Read →](https://www.harperflow.io/never-on-the-podium) [No. 3
Can Your Own Site Get Cited by AI?
p=0.004 Yes — 31% of the time. We tested four theories about why. Three were statistical noise, and we publish them anyway. The one that survived was structure: FAQ blocks and tables.Perplexity only · n=39Read →](https://www.harperflow.io/can-your-site-get-cited-by-ai) [No. 5 · the drift report
The Reddit Collapse
87% → 18% Fifteen days, identical questions, identical pipeline: Reddit fell from 87% of Perplexity's answers to 18%, and the top-cited source changed on 31 of 39 questions — while every structural finding held. The rules are stable; the winners are not.Perplexity only · identical 39Read →](https://www.harperflow.io/the-reddit-collapse) [No. 6
The Webflow Agency AI Visibility Index
2 of 376 Fifty Webflow buyer questions produced 376 named providers — 97% known to exactly one engine, and just two (Finsweet, Flow Ninja) known to all three. Of 1,464 real Webflow agencies, 4.2% were ever named.All 3 enginesRead →](https://www.harperflow.io/webflow-agency-ai-visibility-index)
**The caveat that governs this whole series.** ChatGPT and Gemini do not expose their citation URLs — that's their API, not a gap in our collection. So any finding about *sources* (Reddit, the long tail, page anatomy) is **Perplexity-only, n=39**, and every report says so at each claim. Findings about which *brands* get named can use all three, because names come out of the answer text. Anyone telling you what "AI" cites without naming the engine is guessing.
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Check our work
## The dataset, and the script that proves it
### The 2026 AI Citations Dataset
The raw data behind every report above. Published so you can check our numbers rather than take our word for them — if you recompute something and get a different answer, we want to hear about it.
**queries.csv** · 39 rows · the questions we asked
**multi39.csv** · 117 rows · July 1 baseline
**multi250.csv** · 750 rows · July 16 unified scan (incl. identical 39)
**serp250.csv** · 2,078 rows · Google top-10 per question
**questions250.csv** · 250 rows · the full question set
**reddit-threads.csv** · fetch log (all blocked — disclosed)
**multi39\_enriched.csv** · 39 rows · top-cited pages, coded
**verify-numbers.py** · recomputes every published figure
`python3 verify-numbers.py` recomputes every published figure in all six reports — including the Fisher's exact p-values — straight from those CSVs, and exits non-zero if any of them has drifted. We run it before we publish anything.
It exists because we needed it. On 15 July 2026 we found four figures in our own first edition had gone stale — superseded by a recompute that never made it back into the report, while the old numbers quietly propagated further because they'd had longer to spread. A worked example was wrong too. Prose drifts; the CSV doesn't. Every number is now an assertion against this data. The full correction log is in report No. 1's changelog.
[Download the dataset →](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d6a_harperflow-ai-citations-dataset.zip)
## Does AI recommend you?
Same method, pointed at your domain. We'll ask all three engines your buyers' questions and show you exactly who they name instead of you.
Every report here is produced with the same pipeline we sell — [the one that researches, cites, and publishes](https://www.harperflow.io/features) our own content.
[Run the free audit →](https://www.harperflow.io/geo-website-audit)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/state-of-ai-citations
# We asked three AIs who to buy. They couldn't agree.
ChatGPT, Perplexity, and Gemini now answer the questions your customers used to Google. We put the same 39 buyer questions to all three — 117 answers — and measured exactly who gets recommended, and why.
Prompt asked to all three engines: **"Best project management software for agencies"**
Perplexity
1. 1Productive
2. 2Teamwork.com
3. 3Ravetree
4. 4ClickUp
5. 5Wrike
6. 6Asana
7. 7Monday.com
8. 8Jira
ChatGPT
1. 1Asana
2. 2Trello
3. 3Monday.com
4. 4Basecamp
5. 5Wrike
6. 6ClickUp
7. 7Notion
8. 8Smartsheet
9. 9TeamGantt
10. 10Zenkit
Gemini
1. 1Monday.com
2. 2ClickUp
3. 3Asana
4. 4Teamwork.com
5. 5Wrike
6. 6Adobe Workfront
Three names shared by all three engines, out of fifteen named. *Everything else — different on every engine.*
39 buyer questions × 3 engines = 117 answers · agencies, B2B SaaS, ecommerce · July 2026 · US / English
98%
of "best X" answers named a specific brand
all 3 engines
56%
of questions where all 3 shared even one pick
all 3 engines
\~7%
overlap between the three engines' full lists
all 3 engines
87%
of Perplexity answers cited Reddit
Perplexity only · n=39
**What each number covers.** The first three use all three engines. Everything source-level in this report — Reddit, domain concentration, and the whole page-anatomy section — is **Perplexity only, n=39**, because ChatGPT and Gemini don't expose their citations at all. Nobody can tell you what those two cite, including us.
Finding 01
## AI is a recommendation engine now
Ask any of the three engines a commercial question and it names specific companies almost every time. The era of "ten blue links to choose from" is over — buyers get a short list of names. The only question that matters is whether one of them is yours.
98%
named at least one brand
On commercial "best \[X]" questions. If AI isn't naming you, you're not losing a ranking — you're invisible.
Finding 02 · the big one
## The engines don't agree
Run the same question through all three and they largely recommend **different** companies. All three shared at least one common pick in only 56% of questions. At least two agreed 90% of the time — but their full lists overlapped just \~7%.
"Rank on AI" is really "rank on each AI." You can be ChatGPT's #2 and invisible on Gemini.
How often the engines agreed
Share of the 39 questions, by level of consensus
≥2 engines share a pick
90%
All 3 share a pick
56%
Full-list overlap (avg)
\~7%
0%50%100%
Finding 03
## Perplexity cites Reddit, not your blog
Across every category, Perplexity's most-cited source was Reddit — on 87% of questions — followed by YouTube (38%). Brand-owned blogs and even professional review sites trailed far behind. For at least one major engine, getting cited is less about your blog and more about showing up in the conversations and videos AI already trusts.
Share of Perplexity answers citing each source
% of the 39 questions where the source appeared
Reddit
87%
YouTube
38%
Best review/directory
\~15%
Best brand-owned blog
\~10%
0%50%100%
Source-level data is Perplexity-only — ChatGPT and Gemini don't expose their citations, so we can't yet say whether they lean the same way. Treat this as Perplexity-specific.
Finding 04
## It's a long tail, not winner-take-most
No handful of sites owns AI's citations. The top three domains earned only about a fifth of citations in each category — which means a newcomer can get cited without dethroning an incumbent.
Citation share held by the top 3 domains
By category (Perplexity)
Agencies
18%
B2B SaaS
18%
Ecommerce
21%
0%50%100%
Finding 05
## The cited page: long, listicle, rarely original
The pages AI cites share a shape — long "best-of" listicles, median \~3,200 words. But only about a third contain original data. That last number is the opening: two in three cited pages have **no** original research. It's the emptiest lane in GEO.
What the top cited page had
Share of the 39 top-cited pages (Perplexity)
"Best-of" listicle
77%
Named author
44%
FAQ / tables
38%
Self-published
31%
Original data
31%
0%50%100%
Median cited-page length: **\~3,200 words** · AI's cross-category favorites: Toptal, Finsweet, HubSpot, ActiveCampaign, Flow Ninja.
What to do about it
## The playbook
01
### Win each engine separately
Track your citation share on ChatGPT, Perplexity, and Gemini as three different scoreboards — because they are.
02
### Show up where the AI looks
For Perplexity, that means Reddit threads and YouTube, not just your blog.
03
### Publish what gets cited
Long, genuinely useful comparison and "best-of" pages with clear structure — not thin marketing copy.
04
### Be the 31%
Put original data — your own tests, benchmarks, results — into your content. Two-thirds of cited pages don't, so it's how you get quoted.
Transparency
## Four bugs we caught before publishing
A report is only as good as its worst number. Before publishing we stress-tested our own analysis and fixed four things. We're listing them because a study about *trustworthy content* that hides its own corrections wouldn't deserve to be one.
### What we found and fixed
- **fixed** Our page analysis truncated long pages, undercounting "original data" pages at 5%. The real figure is 31%.
- **fixed** A set-math error reported cross-engine agreement as 0%. The corrected figure was 62%.
- **fixed** Brand extraction initially grabbed client names and technologies as "recommendations." We tightened it.
- **fixed** A final pass recomputed every headline figure straight from the released CSV — and the agreement numbers were *still* wrong. All-three-agree is 56% (not 62%), at-least-two is 90% (not 92%), full-list overlap is \~7% (not \~8%). Those recomputed figures are the ones in this report. We'd already fixed that number once; recomputing from raw data is the only reason we caught it twice.
The data
## Every number, in one place
| Metric | Scope | Value |
| --- | --- | --- |
| Answers named a specific brand | 3 engines | 98% |
| All 3 engines shared ≥1 pick | 3 engines | 56% |
| ≥2 engines shared ≥1 pick | 3 engines | 90% |
| Full-list overlap (avg Jaccard) | 3 engines | \~7% |
| Answers citing Reddit | Perplexity | 87% |
| Answers citing YouTube | Perplexity | \~38% |
| Top-3 domain citation share | Perplexity | 18–21% |
| Cited pages that are listicles | Perplexity | 77% |
| Cited pages with original data | Perplexity | 31% |
| Median cited-page length | Perplexity | \~3,200 words |
| Mean cited-page length (for contrast) | Perplexity | \~4,300 words |
| Questions × engines | total | 39 × 3 = 117 |
**Two notes for anyone recomputing.** **Medians, not means.** The typical cited page is \~3,200 words; the mean is \~4,300, dragged up by a few very long outliers. We report the median because it describes the typical page — if you compute the mean and get 4,300, that's why, not a discrepancy. **The headline survives name-folding.** Engines name the same product differently — Perplexity says "Monday Work Management" where the others say "Monday.com" — and naive string matching would manufacture disagreement. We re-ran everything with 8 such variant groups folded: all-three-agree moved **+0.0 points** (56.4% → 56.4%), overlap +0.2. The disagreement finding is not an artifact of naming.
Changelog
## What changed, and when
This report is versioned. When a number changes, it gets listed here rather than quietly edited.
### v1.1 — 2026-07-15
- **corrected** Four figures had gone stale between an internal recompute and this document: all-three-agree 62% → 56%, at-least-two 92% → 90%, full-list overlap \~8% → \~7%, Perplexity/Reddit "\~9 in 10" → 87%. The dataset never changed; the report had simply not been updated to match it.
- **corrected** The worked example showed Wrike as a Gemini-only pick. All three engines named it — the shared set is Asana, ClickUp *and* Wrike. The side-by-side now reflects what each engine actually returned.
- **corrected** The headline read "117 buyer questions." It is 39 questions × 3 engines = 117 *answers*.
- **added** Scope labels on every headline figure — three of the five are Perplexity-only and the summary didn't say so.
- **added** The name-folding robustness test, the median-vs-mean note, and `verify-numbers.py`, which recomputes all 48 figures in this series from the CSVs and fails if any drifts.
### v1.0 — 2026-07-02
- First edition.
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions this report answers
## Quick answers, straight from the data
### How often do ChatGPT, Perplexity and Gemini agree on a recommendation?
On 39 identical buyer questions in July 2026, all three engines shared at least one pick 56% of the time, at least two agreed 90% of the time, and their full lists overlapped just \~7%.
### Does AI actually name brands in its answers?
Yes. 98% of “best X” answers named at least one specific brand — buyers get a short list of names, not links.
### What sources does Perplexity cite most?
Reddit, in 87% of answers, followed by YouTube at 38% (Perplexity only, n=39 — ChatGPT and Gemini don’t expose their citations).
### What kind of page gets cited by AI?
Long “best-of” listicles — median \~3,200 words, 77% listicles — and only 31% carried original data. Original research is the emptiest lane.
## Does AI recommend you?
Run your brand through the free AI-visibility audit and see exactly what ChatGPT, Perplexity, and Gemini say about you — and whether they name you or your competitor.
If the engines disagree this much about who to cite, the only way to know where you stand is to measure it — you can [run a free GEO audit of your own site](https://www.harperflow.io/geo-website-audit) to see.
[Run my free audit →](https://www.harperflow.io/geo-website-audit) [Download the full dataset](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d6a_harperflow-ai-citations-dataset.zip)
Take this report with you
[Download PDF](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d73_the-2026-state-of-ai-citations-harperflow.pdf) [Download the dataset](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d6a_harperflow-ai-citations-dataset.zip)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/never-on-the-podium
# In every room, never on the podium.
We logged every source Perplexity cited across 39 real buyer questions — all 328 of them. Reddit turned up in 87% of answers, more than twice as often as anything else. It was the top-cited source for exactly none of them.
39 buyer questions · 328 citations · 233 domains · July 2026 · US / English · **Perplexity only — see scope note**
**Data update — July 16, 2026.** We re-ran all 39 questions through the identical pipeline 15 days after collection. **Reddit’s presence fell from 87% of answers to 18%** — the largest single change we have measured — and the top-cited source changed on 31 of 39 questions, while this report’s structural findings (no podium; the long tail) held. This page reports the July 1 snapshot, as labeled; the follow-up drift report covers what changed.
87%
of answers cited Reddit
most of any source
0
questions where Reddit was the top source
out of 39
34
different domains took the top slot
across 39 questions
86%
of cited domains appeared exactly once
200 of 233
**Scope, before anything else.** Every number in this report is **Perplexity, n=39**. ChatGPT and Gemini don't expose their citation URLs at all, so there is no source-level data to analyse for them — not by us, not by anyone. We'd rather publish a narrow true thing than a broad guess.
Finding 01
## Reddit is the most inevitable source on the internet
Reddit was cited in **34 of 39 answers (87%)**. YouTube, in second, managed 15 of 39 (38%). Third place — a single agency site — appeared in 6 of 39, and it falls off a cliff from there.
Nothing else in the data has reach like that. Ask Perplexity a commercial question and a Reddit thread is almost certainly part of the answer. It holds across categories too: ecommerce 92%, B2B SaaS 92%, agencies 77%.
Share of answers citing each source
% of the 39 questions where the source appeared at all · Perplexity only
Reddit
87%
YouTube
38%
Next best domain
15%
0%50%100%
Finding 02 · the one nobody expects
## But Reddit never actually wins
0/ 39
Across all 39 questions, the top-cited page was a Reddit thread **zero times**.
Reddit is cited once per answer, in 87% of answers — **34 of 328 citations, just 10.4% of the total**. It is never the source the engine leans on hardest. It's the source the engine checks itself against.
Reddit isn't the loudest source. It's the most inevitable one.
**Implication:** the standard advice — "go win Reddit" — misreads what Reddit is doing here. It isn't the destination the engine sends you to. It's the sanity check the engine runs before committing to a name. You don't rank on Reddit. You get *mentioned* there, and the engine treats that as corroboration for a recommendation it sources elsewhere.
Finding 03
## There is no podium, only a scramble
Across 39 questions, **34 distinct domains** held the top-cited slot. YouTube took it three times; two domains took it twice; every other winner won exactly once.
Concentration is weak deeper in the data too. The top three domains in each category account for only about a fifth of that category's citations.
Top-3 domain concentration, by category
share of all citations held by that category's three biggest domains
Ecommerce / DTC
21%
Web-design agencies
18%
B2B SaaS
18%
0%50%100%
There is no incumbent to displace. No five-site cartel. The engine reaches for a different source almost every single time.
Finding 04 · the opportunity, as a number
## The tail is the market
**200 of the 233 domains in this study (86%) were cited exactly once.** Reddit and YouTube together — the two most dominant sources in AI citation by a wide margin — account for **15% of all citations**. Everything else, 85% of the pool, is the tail.
The median winner won once. The cost of entry is one page.
That's the whole opportunity stated plainly. The citation graph is not a fortress with a queue outside. It's a wide-open field. **Stop trying to beat the top of the list. There isn't one. Be in the pool.**
What to do about it
## Four things this data actually supports
01
### Earn Reddit mentions, don't chase Reddit rankings
It's corroboration, not traffic. Be the brand people already name in the thread.
02
### Stop benchmarking against a category leader
In this data there isn't one. The top three hold about a fifth, and 86% of domains appear once.
03
### Publish for the tail
One citable page per question you want to win beats one flagship page for the category. Breadth is what the data supports.
04
### Measure per engine
All of this is Perplexity. Nobody can tell you what wins on ChatGPT or Gemini. Be suspicious of anyone who says they can.
The data
## Every number, in one place
| Metric | Scope | Value |
| --- | --- | --- |
| Total citations logged | Perplexity, 39 questions | 328 |
| Distinct domains cited | Perplexity, 39 questions | 233 |
| Mean citations per answer | Perplexity, 39 questions | 8.4 |
| Answers citing Reddit | Perplexity, 39 questions | 34 / 39 = 87% |
| Reddit's share of all citations | Perplexity, 39 questions | 10.4% |
| Questions where Reddit was top-cited | Perplexity, 39 questions | 0 / 39 |
| Answers citing YouTube | Perplexity, 39 questions | 15 / 39 = 38% |
| YouTube's share of all citations | Perplexity, 39 questions | 4.6% |
| Reddit + YouTube share of citations | Perplexity, 39 questions | 15% |
| Domains cited exactly once | Perplexity, 39 questions | 200 / 233 = 86% |
| Distinct top-cited domains | Perplexity, 39 questions | 34 |
| Top-3 domain share — ecommerce | Perplexity, 13 questions | 21% |
| Top-3 domain share — agencies | Perplexity, 13 questions | 18% |
| Top-3 domain share — SaaS | Perplexity, 13 questions | 18% |
**Limits, plainly.** **One engine** — Perplexity, because it's the only one that shows its sources. **n=39 questions, 13 per category**, so category figures are directional. **Domain-level, not thread-level** — we logged `reddit.com`, not which subreddit, so we can't tell you which communities do the work. That's the obvious next study and it's running. **One snapshot, July 2026** — engines change retrieval constantly; we're re-running these questions monthly to measure the drift. The full dataset ships with this report, and `verify-numbers.py` recomputes every figure above straight from it.
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions this report answers
## Quick answers, straight from the data
### How often does Perplexity cite Reddit?
In 34 of 39 answers (87%) in the July 1, 2026 snapshot — more than any other source, across every category we tested.
### Is Reddit the top-cited source in AI answers?
No. Reddit was the top-cited page on 0 of 39 questions, and it accounts for just 10.4% of the 328 citations logged.
### How concentrated are AI citations?
Barely concentrated at all: 86% of the 233 cited domains appeared exactly once, and each category’s top three domains hold only about a fifth of its citations.
## Does AI recommend you?
Same method, pointed at your domain. We'll ask the engines your buyers' questions and show you who they name instead of you.
Winning the long tail means publishing structured, cited answers to hundreds of specific questions — the repetitive work [we built HarperFlow to automate](https://www.harperflow.io/features).
[Run the free audit →](https://www.harperflow.io/geo-website-audit)
Take this report with you
[Download PDF](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d74_in-every-room-never-on-the-podium-harperflow.pdf) [Download the dataset](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d6a_harperflow-ai-citations-dataset.zip)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/can-your-site-get-cited-by-ai
# Can your own site get cited by AI?
Yes — about a third of the time. We set out to document a "self-published penalty" and the data refused to cooperate. So instead we tested four theories about what separates the brand-owned pages that get cited from the third-party pages that do. Three of the four were noise. This report includes them anyway.
39 top-cited pages · 12 self-published vs 27 third-party · Fisher's exact, two-tailed · July 2026 · **Perplexity only — see scope note**
31%
of top-cited pages are the brand's own
12 of 39 — your site is a real lane
1 of 4
theories survived significance testing
the other three are published too
p=0.004
the one that survived: FAQ blocks & tables
75% vs 22%
**Scope, before anything else.** Every number here is **Perplexity, n=39**. ChatGPT and Gemini don't expose citation URLs at all, so there's no page to analyse for them. And n=12 vs n=27 is a small sample — we ran the significance tests precisely because eyeballing percentages at this size is how you fool yourself.
Finding 01
## Your own site is cited about a third of the time
31%
Of 39 top-cited pages, **12 were self-published** — a brand's own page, ranking for a question that brand commercially cares about. **27 were third-party**: listicles, directories, review sites, comparison posts.
So the pessimistic story — "AI won't cite you, only the listicles that outrank you" — is wrong. It cites you roughly a third of the time. The interesting part is what's different about those pages.
Finding 02 · the only one that survived
## Structure was the difference. Nothing else was.
We tested four hypotheses. Here is every result, including the failures:
| Trait | Self-published | Third-party | p (Fisher) | |
| --- | --- | --- | --- | --- |
| **FAQ blocks or data tables** | **75%** | **22%** | **0.004** | significant |
| Contains original data | 42% | 26% | 0.455 | noise |
| Listicle / "best-of" format | 75% | 78% | 1.000 | noise |
| Has a named author | 42% | 44% | 1.000 | noise |
Pages carrying FAQ blocks or data tables
the only trait that separated cited brand pages from cited third-party pages
Self-published (n=12)
75%
Third-party (n=27)
22%
0%50%100%
One trait clears the corrected threshold: **structured content**. Brand-owned pages that get cited are more than 3× as likely to carry an FAQ block or a data table as the third-party pages that get cited.
We think the mechanism is boring and mechanical. An FAQ block is a question with its answer directly underneath. A table is a set of facts with labels attached. Both are trivially extractable. A retrieval engine looking for something quotable finds it faster in a table than in a paragraph. Third-party listicles don't need that help — they win on being a listicle. Your page apparently does.
Structure isn't decoration. It's how the engine finds the quote.
**The honest caveat.** n=12. This is a strong signal in a small sample. We ran four comparisons, so the Bonferroni-corrected threshold is **p < 0.0125**, not 0.05 — this clears it, but we wouldn't bet the company on the exact multiple. We'd bet on the direction.
Finding 03 · the section we'd have cut if we were selling
## Three theories we had to throw away
**Original data didn't separate them.** 42% vs 26% looks like a gap. At n=12 vs n=27 it's p = 0.455 — comfortably inside the range you'd get from coin flips. Elsewhere in our research original data is a live advantage. On *this* question at *this* sample size we can't claim it, so we aren't claiming it.
**Format didn't separate them.** 75% vs 78% listicle. p = 1.000. Cited pages are overwhelmingly listicles regardless of who publishes them. Being a "best-of" post is table stakes, not an edge.
**Bylines didn't separate them.** 42% vs 44%. p = 1.000. There's a great deal of GEO advice built on author/E-E-A-T signals. In this data a named author does nothing to distinguish a cited brand page from a cited third-party page. That's not proof it doesn't matter. It's proof we didn't measure it mattering.
**And length ran the wrong way.** Cited brand-owned pages had a median of **3,158 words**; third-party pages ran **3,901**. If length were the lever, self-published pages would be longer, not shorter.
Finding 04 · the best finding we can least defend
## Category may matter more than anything you do
How often a brand's own page was the top-cited source
by category · 13 questions each · Perplexity only
Web-design agencies
46%
B2B SaaS
38%
Ecommerce / DTC
8%
0%50%100%
An agency's own site was cited nearly half the time. An ecommerce brand's own site was cited **once in thirteen questions**. If that holds at scale it's worth more than every on-page tactic in this report combined.
But we have to be straight with you: agency vs ecommerce is **p = 0.073**. It does not clear significance at 13 questions per category. It's the single most interesting thing in this dataset and the single best argument for running the study bigger. We're running it bigger.
What to do about it
## Four things this data actually supports
01
### Structure your pages
FAQ blocks and data tables are the one thing that separated cited brand pages from cited third-party pages at significance. They're also cheap.
02
### Don't over-invest in bylines on this evidence
They may well matter. They didn't measure as mattering here, and we're not going to pretend otherwise.
03
### Know your category's ceiling
If you sell DTC, your own blog cleared 8% in this sample — third-party placement is likely the higher-leverage play. If you're an agency, your own site is a real lane.
04
### Don't add words
Cited brand pages were *shorter* than cited third-party pages. Length is not the lever.
The data
## Every number, in one place
| Metric | Scope | Value |
| --- | --- | --- |
| Top-cited pages analysed | Perplexity, 39 questions | 39 |
| Self-published | Perplexity, 39 questions | 12 / 39 = 31% |
| Third-party | Perplexity, 39 questions | 27 / 39 = 69% |
| FAQ/tables — self-published | n=12 | 9 = 75% |
| FAQ/tables — third-party | n=27 | 6 = 22% |
| FAQ/tables — significance | Fisher, 2-tailed | p = 0.004 |
| Original data — self-pub vs third-party | n=12 / n=27 | 42% vs 26% · p = 0.455 |
| Listicle — self-pub vs third-party | n=12 / n=27 | 75% vs 78% · p = 1.000 |
| Named author — self-pub vs third-party | n=12 / n=27 | 42% vs 44% · p = 1.000 |
| Median word count — self-published | n=12 | 3,158 |
| Median word count — third-party | n=27 | 3,901 |
| Self-published rate — agencies | n=13 | 6 = 46% |
| Self-published rate — B2B SaaS | n=13 | 5 = 38% |
| Self-published rate — ecommerce | n=13 | 1 = 8% |
| Agencies vs ecommerce | Fisher, 2-tailed | p = 0.073 |
**Limits, plainly.** **One engine** (Perplexity) — the only one that shows its sources. **n=39 pages**, split 12 / 27. One result clears the corrected threshold; the rest are reported as failures on purpose. **13 questions per category**, which is why the most interesting finding is the one we can least defend. **Coding is judgement** — "has original data" and "self-published" were coded by rule, but edge cases exist. The dataset ships with this report so you can disagree with our specific calls rather than in general. **One snapshot, July 2026.** And `verify-numbers.py` recomputes every figure above straight from the CSVs.
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions this report answers
## Quick answers, straight from the data
### Can a brand’s own website get cited by AI?
Yes — 12 of the 39 top-cited pages (31%) were the brand’s own page, not a third-party listicle (Perplexity, July 2026).
### What makes a brand page citable?
Structure. Cited brand pages carried FAQ blocks or data tables 75% of the time versus 22% for cited third-party pages — p=0.004, the only significant trait of the four we tested.
### Do bylines or word count help a page get cited?
Not measurably here. Named authors: 42% vs 44% (p=1.000). And cited brand pages were shorter than third-party ones — median 3,158 vs 3,901 words — so length is not the lever.
## Is your own site getting cited?
Same method, pointed at your domain. We'll ask the engines your buyers' questions and show you whether your pages — or someone else's — come back.
Structuring content this way — FAQ blocks, tables, clear answers — is exactly what [HarperFlow's publishing pipeline](https://www.harperflow.io/features) does on every article.
[Run the free audit →](https://www.harperflow.io/geo-website-audit)
Take this report with you
[Download PDF](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d75_can-your-own-site-get-cited-by-ai-harperflow.pdf) [Download the dataset](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d6a_harperflow-ai-citations-dataset.zip)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/one-engines-favourite
# One engine's favourite. Everyone else's unknown.
The most-recommended brand across our 39 buyer questions is **Toptal** — 10 mentions, more than any other company. Every single one came from ChatGPT. Perplexity and Gemini have never heard of it. And Toptal is not the exception: **79% of brands exist in exactly one engine's world.**
39 buyer questions · 117 answers · 646 distinct brands · July 2026 · US / English · **all three engines**
646
distinct brands named across the study
all 3 engines
79%
of brands are named by exactly one engine
509 of 646
8%
of brands are named by all three
just 51 companies
1 of 3
engines have ever named the #1 brand
ChatGPT only
**A rare one: no engine caveat.** Most of this series is Perplexity-only, because ChatGPT and Gemini don't expose their citations. This report reads brand *names* out of the answer text — which all three engines give us — so every figure here uses all three.
Finding 01
## Toptal: number one, and two-thirds invisible
**Toptal was the most-named brand in the study** — 10 mentions across 117 answers, more than any other company. It is also, as far as Perplexity and Gemini are concerned, not a thing that exists.
ChatGPT
10
mentions
Perplexity
0
mentions
Gemini
0
mentions
Sit with what that means. If you benchmark your AI visibility against Toptal — the category's apparent AI darling — and you only ever check ChatGPT, you'd conclude they have solved this. Check Perplexity and they're nowhere. **Two-thirds of the market cannot see the market leader.**
You cannot know how you're doing from one engine.
Not directionally. Not roughly. The #1 brand in this study scores 10 on one engine and 0 on the others.
Finding 02 · the normal case
## 79% of brands exist in exactly one engine's world
Toptal is not a quirk. It's the normal case, drawn in bold.
How many engines know each brand?
646 distinct brands · all 3 engines · 39 questions
Exactly 1 engine
79%
Exactly 2 engines
13%
All 3 engines
8%
0%50%100%
Of 646 brands, only **51 are known to all three engines**. Four out of five exist inside exactly one engine's answers.
This is the "three separate games" finding from our first report, restated at the level that actually matters to you. It isn't just that the engines return different *lists*. It's that the overwhelming majority of companies in this market are visible to exactly one of the three — and because none of this shows up in analytics, almost none of them know which.
Finding 03 · the twist
## Fame is shared. Obscurity is fragmented.
Here's what stops this being a doom story. Look at the 20 most-named brands and the pattern inverts: **18 of the top 20 are known to two or three engines.**
Finsweet (9 mentions), HubSpot (7), ActiveCampaign (7), Flow Ninja (6), Pipedrive (6), Zoho CRM (6), Freshdesk (6), Pendo (6) — all three engines name every one of them. Only Toptal and one other are single-engine acts.
So the market has two regimes:
The top
### 51 brands · 8%
A small shared set all three engines have converged on. Enough presence that every engine's retrieval finds them.
The tail
### 509 brands · 79%
Each living inside exactly one engine's answer set, invisible to the other two.
The gap between the tail and the top isn't fame. It's coverage.
**The implication is not "you can't win."** It's that cross-engine presence is the thing that separates the two regimes — and it's a *reachable* state, because the shared set is 51 companies deep, not five.
Finding 04
## The engines have different appetites
They don't even name the same *number* of companies.
| Engine | Brands per answer (mean) | Range | Distinct brands |
| --- | --- | --- | --- |
| ChatGPT | 8.9 | 0–13 | 296 |
| Perplexity | 8.3 | 2–18 | 292 |
| Gemini | 7.2 | 0–23 | 246 |
ChatGPT is the most consistent recommender — it names the most brands on average and never went above 13. Gemini is the most volatile: it named nothing at all on some questions and 23 companies on another. Perplexity never returned an empty answer.
A practical read: Gemini's shortlists are shorter on average, which makes each slot scarcer and being absent more costly.
What to do about it
## Four things this data actually supports
01
### Never measure on one engine
The most-recommended brand in this study scores 10 on ChatGPT and 0 elsewhere. One engine tells you nothing about the other two.
02
### Aim for "named by all three"
Not "ranked #1 on one." Only 8% of brands manage it — that's the moat, and it's 51 companies wide, not five.
03
### Find out which engine sees you
If you're in the 79%, you don't know. Optimising for the engine that already names you is the most common wasted effort in this category.
04
### Don't copy the visible leader
They may be a one-engine act, and you'd be copying a strategy two-thirds of the market can't see.
The data
## Every number, in one place
| Metric | Scope | Value |
| --- | --- | --- |
| Distinct brands named | all 3 engines, 39 questions | 646 |
| Named by exactly 1 engine | all 3 engines | 509 = 79% |
| Named by exactly 2 engines | all 3 engines | 86 = 13% |
| Named by all 3 engines | all 3 engines | 51 = 8% |
| Most-named brand | all 3 engines | Toptal · 10 mentions |
| Engines naming Toptal | all 3 engines | 1 of 3 (ChatGPT) |
| Of the top 20, single-engine | all 3 engines | 2 of 20 |
| Brands per answer — ChatGPT | 39 answers | mean 8.9 · range 0–13 |
| Brands per answer — Perplexity | 39 answers | mean 8.3 · range 2–18 |
| Brands per answer — Gemini | 39 answers | mean 7.2 · range 0–23 |
| Distinct brands — ChatGPT | 39 answers | 296 |
| Distinct brands — Perplexity | 39 answers | 292 |
| Distinct brands — Gemini | 39 answers | 246 |
**Limits, plainly.** **n=39 questions, 117 answers, one snapshot (July 2026).** A brand named once by one engine counts as "named by one engine" — at this sample size there is real noise in the tail, and the 79% will move with n. The direction is robust; the decimal isn't. **"Named by" is a low bar** — one mention in one answer. This measures presence, not prominence. **Naming variants are unfolded** — "Monday Work Management" and "Monday.com" count separately, which would slightly *inflate* the single-engine share. Folding the 8 variant groups we found moves the series' headline agreement figure by +0.0 points. **Model versions move** — sonar / gpt-4o-mini / gemini-2.5-flash, July 2026. The full dataset ships with this report, and `verify-numbers.py` recomputes every figure above straight from it.
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions this report answers
## Quick answers, straight from the data
### Do the AI engines recommend the same brands?
Mostly not. Of 646 brands named across 117 answers, 79% were named by exactly one engine; only 51 brands (8%) are known to all three.
### What was the most-recommended brand — and who recommends it?
Toptal, with 10 mentions — every single one from ChatGPT. Perplexity and Gemini never named it once.
### How many brands does each engine name per answer?
ChatGPT a mean of 8.9 (range 0–13), Perplexity 8.3 (2–18, never empty), Gemini 7.2 (0–23, the most volatile).
## Which engines can see you?
Same method, pointed at your domain. We'll ask all three engines your buyers' questions and show you which ones name you — and which ones name someone else.
Because visibility varies so much by engine, it's worth checking how each sees you — [a free GEO audit](https://www.harperflow.io/geo-website-audit) shows where you surface and where you're invisible.
[Run the free audit →](https://www.harperflow.io/geo-website-audit)
Take this report with you
[Download PDF](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d76_one-engines-favourite-harperflow.pdf) [Download the dataset](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d6a_harperflow-ai-citations-dataset.zip)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/the-reddit-collapse
# The Reddit collapse.
Fifteen days ago, Reddit appeared in 87% of Perplexity's answers to our 39 buyer questions — the most inevitable source we had ever measured. We re-ran the **identical questions** through the **identical pipeline**. It's 18%.
87%→18% answers citing Reddit · byte-identical 39 questions · July 1 → July 16, 2026 · Perplexity
39 identical questions · 15 days apart · same engines, prompts, gateway, extraction · **source-level figures are Perplexity only**
79%
of questions changed their single top-cited source
31 of 39 · Perplexity
25%
Perplexity's brand-list overlap with its own answers from 15 days ago
ChatGPT 37% · Gemini 38%
59%
all-3-engine agreement — structurally unchanged
was 56% · all 3 engines
2.0%
Reddit's share of all citations now
was 10.4%
**How controlled is this comparison?** The 39 questions are byte-identical (mechanically enforced), same three engines and model ids, same system prompt and locale, same gateway, same counting method as report No. 2. One honest caveat: both runs pass through the same API gateway, so we measure the identical *pipeline's* behavior — we can't fully apportion the change between the engine and its citation passthrough. Either way, nothing on our side changed.
Finding 01
## The collapse — with its controls
Only seven questions cited Reddit both times. And the two controls say this is Reddit-specific, not a citation drought: answers actually got **richer** (8.4 → 9.1 cited domains per answer), and YouTube barely moved.
| Set A — identical 39 questions | July 1 | July 16 |
| --- | --- | --- |
| **Answers citing Reddit** | 34 / 39 = 87% | **7 / 39 = 18%** |
| Reddit's share of all citations | 10.4% | 2.0% |
| Answers citing YouTube (control) | 15 / 39 = 38% | 14 / 39 = 36% |
| Cited domains per answer (control) | 8.4 | 9.1 |
Finding 02
## Where the citations went
The hole Reddit left was filled — largely by editorial and listicle publishers. Community corroboration was swapped for edited, publisher-shaped pages.
Most-cited domains on the identical 39 questions
citation counts, Set A · Perplexity only
reddit.com — July 1
34
reddit.com — July 16
7
youtube.com — July 16
14
forbes.com — July 16
7
designrush.com — July 16
6
goodhousekeeping.com — July 16
6
0bars scaled to 39 questions39
If that direction holds, the pool got *more* winnable by published pages — which is most of what a brand can actually control — and less dependent on being talked about in threads.
Finding 03
## It isn't just Reddit. Everything churns.
The single top-cited source changed on **31 of 39 questions (79%)** in fifteen days. And at the brand level, each engine barely agrees with *itself* across the fortnight:
| Engine | Overlap with its own July 1 brand list |
| --- | --- |
| Perplexity | **25%** |
| ChatGPT | 37% |
| Gemini | 38% |
| For scale: two different engines, same day | 7.5% |
"Your brand ranks on ChatGPT" is a statement with a half-life measured in days.
Finding 04 · the counterweight
## The rules are stable. The winners are not.
While the contents shuffled wholesale, the *shape* of the system did not move:
| Structural metric — Set A | July 1 | July 16 |
| --- | --- | --- |
| All 3 engines shared ≥1 pick | 56% | 59% |
| At least 2 engines shared ≥1 pick | 90% | 87% |
| Full-list overlap (mean Jaccard) | 7.2% | 7.5% |
Every structural claim this series has published survived the fortnight. Every source-specific fact was perishable. That asymmetry *is* the strategy: **optimize for the structure — be citable, be structured, be in the pool — not for any single incumbent source.**
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
What remains
## Reddit today, across all 250 questions
In the full July 16 scan (250 questions), Reddit appears in **13% of answers** and holds **1.8% of all citations** — 36 thread URLs spread across 30 different subreddits, none appearing more than three times. The subreddit taxonomy we had planned is therefore moot: there is no concentration left to map.
**Disclosure.** We attempted to fetch the 35 threads' public metadata for age analysis; reddit.com returned HTTP errors for every unauthenticated request. We don't scrape around blocks, so thread ages are not reported — degraded scope over fabricated scope. And a limit worth repeating: **two snapshots make a line, not a law.** This is the first interval of a monthly series; Reddit could be back at 87% next month, and that would be just as publishable.
What to do about it
## Four things this data actually supports
01
### Date every source-specific stat
"Reddit is 87% of AI answers" was true, measured — and lasted under fifteen days. Ask anyone quoting a citation stat when it was measured, and how often they re-measure.
02
### Optimize for the structure, not the incumbent
The durable facts are structural: engines disagree, agreement is shallow, the pool is a long tail. Structured, citable pages on your own domain competed in both versions of this landscape; a Reddit-thread strategy competed in exactly one.
03
### Measure continuously or not at all
Monthly is now the minimum honest cadence. This series re-runs the same 39 questions every month and publishes whatever they say.
04
### Re-read report No. 2 with today's numbers
Its structural findings held; its Reddit reach figure did not — which is why it carries a dated update banner instead of a quiet edit.
The data
## Every number, in one place
| Metric | Scope | Value |
| --- | --- | --- |
| Reddit presence — July 1 | Perplexity, Set A n=39 | 34 / 39 = 87% |
| Reddit presence — July 16 | Perplexity, Set A n=39 | 7 / 39 = 18% |
| Questions citing Reddit both runs | Perplexity, Set A | 7 |
| Reddit share of citations — July 1 | Perplexity, Set A | 10.4% |
| Reddit share of citations — July 16 | Perplexity, Set A | 2.0% |
| YouTube presence, July 1 → 16 | Perplexity, Set A | 38% → 36% |
| Cited domains per answer | Perplexity, Set A | 8.4 → 9.1 |
| Top-cited domain changed | Perplexity, Set A | 31 / 39 = 79% |
| Self-overlap over 15 days — Perplexity | brand lists, n=39 | 25% |
| Self-overlap — ChatGPT / Gemini | brand lists, n=39 | 37% / 38% |
| Cross-engine overlap, same day | all 3, July 16 | 7.5% |
| All-3 agreement, July 1 → 16 | all 3, Set A | 56% → 59% |
| ≥2 agreement, July 1 → 16 | all 3, Set A | 90% → 87% |
| Reddit presence, all 250 questions | Perplexity, July 16 | 33 / 250 = 13% |
| Reddit share of all citations | Perplexity, July 16 n=250 | 1.8% |
| Distinct subreddits / thread URLs | Perplexity, July 16 | 30 / 36 |
| Thread metadata fetches blocked | reddit.com, unauthenticated | 35 / 35 |
**Recompute us.** Both datasets ship with this report — `multi39.csv` (July 1) and `multi250.csv` (July 16; Set A is the identical re-run) — and `verify-numbers.py` asserts every figure above straight from them. We don't know *why* the collapse happened — licensing, crawl access, retrieval re-ranking are all consistent with the data and none is verifiable from outside — so we report the measurement and decline to speculate.
Questions this report answers
## Quick answers, straight from the data
### How fast do AI answers actually change?
In 15 days, on byte-identical questions and an identical pipeline, Reddit fell from 87% of Perplexity’s answers to 18%, and the top-cited source changed on 31 of 39 questions (79%).
### Did the structure of AI answers change too?
No. All-three-engine agreement went 56% → 59%, at-least-two went 90% → 87%, and full-list overlap went 7.2% → 7.5%. The rules held; the winners churned.
### Is the churn just a Perplexity problem?
No engine agrees with its own two-week-old answers: brand-list overlap with July 1 was 25% for Perplexity, 37% for ChatGPT, 38% for Gemini.
## Which sources cite you — this week?
Same method, pointed at your domain, re-measurable monthly. We'll ask the engines your buyers' questions and show you what they cite instead of you — with a date on it.
It's the same lesson we built HarperFlow around: in AI search, what you publish matters more than what you score — which is why we pitch it as [a Surfer SEO alternative that publishes, not just grades](https://www.harperflow.io/surferseo-alternative).
[Run the free audit →](https://www.harperflow.io/geo-website-audit)
Take this report with you
[Download PDF](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d77_the-reddit-collapse-harperflow.pdf) [Download the dataset](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d6a_harperflow-ai-citations-dataset.zip)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/webflow-agency-ai-visibility-index
# The Webflow Agency AI Visibility Index
We asked ChatGPT, Perplexity and Gemini fifty questions a real Webflow buyer asks — by service, industry, geography and company size. They named **376 providers**. Exactly **two** are known to all three engines.
50 Webflow buyer questions · 150 answers · July 16, 2026 · US / English · **brand names come from answer text — all three engines count**
376
providers named across 150 answers
821 mentions
97%
of providers are known to exactly one engine
365 of 376
2
providers known to all three engines
Finsweet · Flow Ninja
4.2%
of 1,464 real Webflow agencies were ever named
internal roster · a floor
**What was filtered, by rule.** Before counting we removed any "provider" whose name contains the word *webflow* ("Webflow Experts directory", "the Webflow community" — Webflow's brand terms bar agencies from using the name, so these are references, not companies) plus talent marketplaces (Toptal, Upwork, Fiverr, Clutch, Contra, Dribbble, Behance): 24 entries, 91 mentions, full list in the dataset. Name variants are not folded, so counts are conservative floors. **The Index measures presence in AI answers, not quality** — no provider here has any commercial relationship with HarperFlow.
Finding 01
## The Index
| # | Provider | Mentions | Engines | |
| --- | --- | --- | --- | --- |
| 1 | **Finsweet** | 90 | 3 / 3 | all-three club |
| 2 | **Flow Ninja** | 62 | 3 / 3 | all-three club |
| 3 | Edgar Allan | 39 | 2 / 3 | |
| 4 | Refokus | 39 | 2 / 3 | |
| 5 | Flowbase | 20 | 1 / 3 | templates — not an agency |
| 6 | Relume | 18 | 1 / 3 | components — not an agency |
| 7 | Veza Digital | 14 | 1 / 3 | |
| 8 | Flowout | 12 | 2 / 3 | |
| 9 | BX Studio | 10 | 1 / 3 | |
| 9 | 8020 | 10 | 2 / 3 | |
| 9 | Amply | 10 | 1 / 3 | |
| 9 | Shadow Digital | 10 | 1 / 3 | |
Finsweet's lead is enormous — 90 mentions is more than positions 5 through 12 combined. And note rows 5 and 6: asked for an *agency*, the engines put two *products* in the top six. They don't reliably distinguish "who can build this for me" from "what tool exists."
Finding 02
## The two-name club
How many engines know each provider?
376 providers · 50 questions · all 3 engines
Exactly 1 engine
97%
Exactly 2 engines
9 providers
All 3 engines
2 providers
0%50%100%
Our general-market study found 79% of brands were single-engine. We expected a tight niche to consolidate — a small market where the engines converge on one short list. The opposite: **97%**. Each engine has assembled its own, almost entirely private, map of the Webflow market.
If your prospect asks ChatGPT and their colleague asks Perplexity, they hear about different companies.
Unless you're one of two names, at most one of those conversations includes you.
Finding 03
## Engines agree *less* in the niche, not more
Questions where all three engines shared ≥1 recommendation
same day, same engines — July 16, 2026
General questions (n=39)
59%
Webflow questions (n=50)
30%
0%50%100%
At least two engines agreed on 80% of niche questions — but full three-way consensus halves when the questions get specialized. Niche depth doesn't create consensus; it removes the universally famous names that produce it. The "three separate games" finding is at its most extreme exactly where specialized businesses live.
Finding 04 · the one that matters
## For 24 of 25 agencies, this is an existence problem
4.2%
We track **1,464 real Webflow-focused agencies** (a proprietary roster built from partner directories and public rosters, matched by normalized name and domain root). Across all 150 answers to all 50 questions, **61 of them were named even once, by any engine.**
There is a head — five names take 30% of all mentions — but the defining feature of this market's AI layer is *absence*: 95.8% of the real market is invisible to it. The gap between position 12 on the Index (10 mentions) and the median tracked agency (zero mentions, zero engines) is the entire game.
**Label.** The 4.2% is the one figure readers can't recompute from the public dataset (the roster is proprietary), and name/domain matching misses renames — treat it as a floor. Everything else on this page recomputes from `multi250.csv`, Set C.
What to do about it — if you're a Webflow agency
## Four things this data actually supports
01
### Find out which engines know you exist
Before optimizing anything. Ask all three your own buyers' questions — given the 97% finding, one engine's answer tells you nothing about the other two.
02
### Copy the club's mechanism, not its fame
Finsweet and Flow Ninja publish prolific, structured, genuinely useful material on their own domains — the exact page profile this series keeps finding in citations. That's a content posture, not a budget.
03
### Don't fight the tools for mentions
Flowbase and Relume rank because engines conflate products with providers. Competing with component libraries for "best agency" mentions is the wrong battle.
04
### Treat it as a standing index, not a verdict
Report No. 5 measured fortnight-scale churn in these same answers. A snapshot's losers are next month's entrants — we'll re-run it.
The data
## Every number, in one place
| Metric | Scope | Value |
| --- | --- | --- |
| Questions / answers | Set C, all 3 engines | 50 / 150 |
| Named providers (after filter) | all 3 engines | 376 |
| Total mentions (after filter) | all 3 engines | 821 |
| Generic/marketplace entries removed | filter rule | 24 entries / 91 mentions |
| Single-engine providers | all 3 engines | 365 / 376 = 97% |
| Two-engine providers | all 3 engines | 9 |
| Known to all three | all 3 engines | 2 |
| Top-5 share of mentions | all 3 engines | 250 / 821 = 30% |
| Niche all-3 agreement | Set C | 15 / 50 = 30% |
| Niche ≥2 agreement | Set C | 40 / 50 = 80% |
| General all-3 agreement, same day | Set A | 59% |
| Tracked agencies ever named | internal roster, n=1,464 | 61 = 4.2% |
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Questions this report answers
## Quick answers, straight from the data
### Which Webflow agencies do AI engines actually recommend?
Only two providers were named by all three engines across 50 Webflow buyer questions: Finsweet (90 mentions) and Flow Ninja (62). 97% of the 376 named providers are known to just one engine.
### How many Webflow agencies are visible to AI at all?
61 of the 1,464 agencies we track (4.2%) were named even once by any engine — and that figure is a floor. For most of the market this is an existence problem, not a ranking problem.
### Do the engines agree more on niche questions?
Less. All-three agreement was 30% on the Webflow questions versus 59% on general questions asked the same day — specialization removes the famous names that create consensus.
## Do the engines know you exist?
Same method, pointed at your agency. We'll ask all three engines your buyers' questions and show you which ones name you — and which name someone else.
For Webflow teams, closing that visibility gap starts with the fundamentals in our [complete guide to Webflow SEO](https://www.harperflow.io/webflow-seo).
[Run the free audit →](https://www.harperflow.io/geo-website-audit)
Take this report with you
[Download PDF](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d79_webflow-agency-ai-visibility-index-harperflow.pdf) [Download the dataset](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a68c4fd0842e5f8fa5b2d6a_harperflow-ai-citations-dataset.zip)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/dead-internet
# Six corners of the web AI is switching off.
Famous corners of the internet are going quiet as AI answers the questions, makes the images, and does the small jobs people used to come for. This is a running series that measures those collapses — each one anchored to public, recomputable data, each number something you can pull yourself. No estimates, no vibes.
## Every figure recomputable
Each number here is a query away — a public API, a stock ticker, a pageviews feed. Pull it and check us.
## One honest window, not the scariest one
Stocks carry bubbles, SPACs, a failed merger; Quora is private. We fence off every confounder and anchor each collapse to its cleanest defensible AI-era number.
## We cut what didn't hold
Digg and Tumblr were on the list. The data said their declines weren't AI — Digg's interest even grew — so they aren't here.
The series
## Six collapses, one cause
Five domains of the human web — developer Q&A, homework, stock imagery, freelance work, general Q&A — plus the giant that sued. Each report measures one collapse from a source you can re-query.
[No. 1
The Great Stack Overflow Collapse
−98.5% New developer questions per month fell from 109,301 the month ChatGPT launched to 1,624 — a 98.5% collapse in the human Q&A that trained the machines.Stack Exchange API · 127 monthsRead →](https://www.harperflow.io/dead-internet/stack-overflow-collapse) [No. 2
The Chegg Collapse
−48% in a day The homework giant lost half its value the morning after it named ChatGPT on an earnings call — $96.53 down to 98 cents, below the delisting line.Yahoo Finance · monthly closeRead →](https://www.harperflow.io/dead-internet/chegg-collapse) [No. 3
The Shutterstock Collapse
$121 → $5 It licensed its library to OpenAI and shipped its own generator — and generative image AI commoditized stock photography anyway.Yahoo Finance · WikimediaRead →](https://www.harperflow.io/dead-internet/shutterstock-collapse) [No. 4
The Fiverr Collapse
the two-year fuse Gig work didn't crash on launch day. Fiverr held flat for two years, then fell 71% as the models finally got good enough to do the tasks.Yahoo Finance · monthly closeRead →](https://www.harperflow.io/dead-internet/fiverr-collapse) [No. 5
The Quora Collapse
−43% The place you went to ask a human. Public search interest fell 43% as ChatGPT began answering the exact questions Quora was built for.Wikimedia pageviews · interestRead →](https://www.harperflow.io/dead-internet/quora-collapse) [No. 6
The Getty Images Collapse
sued & lost Getty took Stability AI to court in two countries — and lost the case and 98.6% of its value. You can't litigate your way out of a capability shift.Yahoo Finance · court recordRead →](https://www.harperflow.io/dead-internet/getty-images-collapse)
**The honest fine print.** Three of these are share prices, which carry more than AI: we throw away the 2021 bubble peaks, flag Getty's SPAC debut, and fence off the failed Getty–Shutterstock merger, anchoring each to the cleanest pre-confounder window. Quora is privately held, so we measure public search interest, not traffic, and say so at every turn. Where a number could be read two ways, each report shows both and tells you which one we lead with.
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
Check our work
## Every number, from a public source you can query
### No dashboard. No private panel.
You don't need our spreadsheet — you need the same open feeds we used. Every figure in the series comes from one of three public sources, and each report names which one it leans on.
**Stack Exchange API** · new questions per month · Part 1
**Yahoo Finance** · monthly closing price · Parts 2 · 3 · 4 · 6
**Wikimedia Pageviews API** · monthly article pageviews · Part 5
Pull the Stack Exchange question total for the `stackoverflow` tag month by month; hit the Yahoo Finance monthly-chart endpoint for CHGG, SSTK, FVRR and GETY; query the Wikimedia pageviews API for the Quora article. Same inputs, same numbers. If you recompute a figure and get a different answer, we want to hear about it.
Prefer not to pull them yourself? Download the full bundle — six CSVs, a README, and `verify.py`, which re-pulls the live sources and asserts every dated anchor figure, exiting non-zero if anything has drifted.
[Download the dataset →](https://cdn.prod.website-files.com/6a68c4fd0842e5f8fa5b2a5e/6a6fee0ab488a89f12533060_dead-internet-report-dataset.zip)
## Does AI answer for you — or instead of you?
Every site in this series was a place people went for something AI now provides directly. The businesses that survive the shift are the ones AI answers cite and send people to. Run a free GEO audit and find out which side of that line you're on.
The takeaway for anyone publishing is the one behind [how we think about AI and Webflow content](https://www.harperflow.io/webflow-ai): generation is cheap, but only cited, checked work survives.
[Run the free audit →](https://www.harperflow.io/geo-website-audit)
Related from HarperFlow
- [How to get cited by AI](https://www.harperflow.io/how-to-get-cited-by-ai)— what ChatGPT, Perplexity and Google actually quote, and the moves that help.
- [The Webflow Agency AI Visibility Index](https://www.harperflow.io/webflow-agency-ai-visibility-index)— which agencies AI names most, measured across 50 buyer questions and three engines.
- [Soro alternative](https://www.harperflow.io/soro-alternative)— an AI publisher built to be cited, with an independent audit and a human veto.
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/dead-internet/chegg-collapse
# The Chegg collapse.
Chegg spent the pandemic as the internet's homework machine — a $12 billion company students paid $20 a month to answer their questions. Then those questions became free, instant, and infinite. Its stock is now worth **98 cents**.
$96.53→$0.98 share price · Feb 2021 peak → Jul 2026 · Yahoo Finance, monthly close
153 months of price history · Dec 2013 → Jul 2026 · monthly closing price · public market data · **every figure recomputable**
−99%
from its pandemic-era peak
$96.53 (Feb 2021) → $0.98
−48%
in a single day — May 2, 2023
the morning after it named ChatGPT
−50%
across the month of that admission
Apr $17.98 → May $8.98, 2023
$0.98
share price now — below the $1 delisting line
Jul 2026
**Isn't a share price just sentiment?** Partly — but it is the market's live estimate of future cash, and it moved 48% in a day on a single sentence about AI. The direction is corroborated by Chegg's own filings: falling subscribers, falling revenue, and repeated layoffs across 2023–2025. Price is the cleanest public, recomputable signal; the fundamentals point the same way.
Finding 01
## A pandemic darling, then a straight line down.
Chegg rode remote schooling from about $35 to a $96.53 peak in February 2021 — briefly a $12 billion company. What followed wasn't a correction. It was the market re-rating the entire business the moment students got a free tutor that never sleeps. Every point below is a monthly close.
{"start":"2019-01","unit":"$","values":\[35.22,39.63,38.12,35.65,37.46,38.59,44.92,39.64,29.95,30.66,38.77,37.91,41.23,39.21,35.78,42.75,61.08,67.26,80.97,73.74,71.44,73.44,77.93,90.33,95.26,96.53,85.66,90.33,76.91,83.11,88.63,83.22,68.02,59.44,27.85,30.7,26.47,31.27,36.28,24.74,19.46,18.78,21.3,19.68,21.07,21.57,29.84,25.27,20.76,15.89,16.3,17.98,8.98,8.88,10.13,10.21,8.92,7.53,9.93,11.36,9.85,8.94,7.57,5.17,3.83,3.16,3.41,2.16,1.77,1.6,2.11,1.61,1.54,1.09,0.64,0.73,1.02,1.21,1.31,1.47,1.51,1.06,0.97,0.93,0.77,0.66,0.74,1.09,1.31,1.01,0.98,0.98],"markers":\[{"month":"2023-05","label":"ChatGPT admission"}],"endLabel":"$0.98"}
Finding 02
## The day it said the word “ChatGPT.”
On May 1, 2023, Chegg's CEO told investors that ChatGPT was starting to dent new-customer growth — among the first times a public company blamed a generative-AI product for its numbers in plain language. The market re-priced overnight.
| Date | Close | What happened |
| --- | --- | --- |
| Feb 2021 | $96.53 | Pandemic peak |
| Apr 2023 | $17.98 | Already down 81% from peak |
| May 2023 | $8.98 | “ChatGPT” admission · −50% in a month |
| Jul 2026 | $0.98 | Sub-dollar · delisting territory |
A single sentence about AI erased half a public company in a day.
Finding 03
## Homework was the perfect thing for AI to eat.
Chegg's product was a paywall around answers — $15–20 a month to see a worked solution a human had once written. That is exactly the shape of task a language model does for free, instantly, with no subscription and no wait. When the thing you sell is *the answer*, and answers become abundant, there is no moat left to defend. The textbook-rental giant didn't lose to a competitor; it lost to a feature.
What it means for you
## If your product is “the answer,” AI is your competitor.
Chegg is a warning to every business whose value is answering a question a model can now answer for free. The subscription middleman between a person and an answer is the most exposed position on the internet. Three lessons.
**1 · Own the answer, don't gate it.** Value locked behind a paywall an AI can replicate for free evaporates. Value from proprietary data, workflow, or trust survives.
**2 · Be the source the model cites, not the site it replaces.** If AI answers from your content and credits you, disintermediation becomes distribution. If it answers instead of you, you're Chegg.
**3 · Watch the tell.** Chegg's stock moved months before its revenue did. The market prices the AI threat before the income statement admits it — and so do your competitors' customers.
The data
## Every figure, one table.
| Metric | Value |
| --- | --- |
| Peak monthly close (Feb 2021) | $96.53 |
| Latest monthly close (Jul 2026) | $0.98 |
| Change from peak | −99.0% |
| Single-day drop, May 2, 2023 | −48% |
| Month of the ChatGPT admission | $17.98 → $8.98 (−50%) |
| Months of history measured | 153 (Dec 2013 – Jul 2026) |
| Source | Yahoo Finance, monthly close |
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
FAQ
## Questions, answered.
**Was it really ChatGPT, or just a frothy stock falling?** Both — but the timing is unambiguous: the largest single-day move in Chegg's history came the day after it named ChatGPT on an earnings call, May 2, 2023. Markets had months of macro to sell on before that; they waited for the AI sentence.
**Is the company failing, or just the stock?** Filings show falling subscribers, falling revenue, and repeated layoffs across 2023–2025. The stock led; the fundamentals followed it down.
**Can I reproduce these numbers?** Yes — every figure is monthly closing price from public market data (Yahoo Finance). Pull the ticker and check.
## Is AI a distribution channel for you — or a replacement?
The businesses that survive the shift are the ones AI answers cite, not the ones it makes redundant. Run a free GEO audit and find out which side you're on.
The properties that survive AI produce genuinely useful, sourced material — the standard [HarperFlow holds every article to](https://www.harperflow.io/features).
[Run the free audit](https://www.harperflow.io/geo-website-audit)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/dead-internet/fiverr-collapse
# The Fiverr collapse.
Fiverr built a marketplace for the internet's small jobs — a logo, a blog post, a voiceover, five dollars and up. The pitch was that any task could be handed to a human somewhere cheaper. Then a chatbot started doing those same tasks for free, in seconds. The stock didn't crash right away. It waited two years — then fell **71%**.
$30.95→$8.93 share price · Oct 2022, the month before ChatGPT → Jul 2026 · Yahoo Finance, monthly close
85 months of price history · Jul 2019 → Jul 2026 · monthly closing price · public market data · **every figure recomputable**
−71%
since the month before ChatGPT
$30.95 (Oct 2022) → $8.93
2 years
the fuse — flat, then the cliff
\~$31 held through Dec 2024
−73%
public search interest in Fiverr
2022 → 2025 avg · Wikimedia pageviews
−84%
Upwork, the other gig giant
fell as far before a partial bounce
**Isn't this just the 2021 bubble deflating?** That's exactly why we throw the peak away. Fiverr hit $270 in the 2021 remote-work mania; by October 2022 — the month before ChatGPT — the bubble was already gone and the stock sat at $30.95. Everything we measure starts there. From that sober post-bubble base, with no mania left to unwind, it still lost 71%. And the shape is the tell: it didn't fall when ChatGPT launched — it held for two years, then broke as the models got good enough to do the work.
Finding 01
## A marketplace for tasks a machine now does for free.
Fiverr's promise was arbitrage: any small job — copywriting, translation, a quick logo, a voiceover — handed to a human somewhere who'd do it cheaper. It rode remote work to a $270 peak in early 2021, then gave the whole bubble back. The interesting part isn't the mania. It's what the stock did from its sober 2022 base, once a language model could do those same small jobs for nothing. Every point below is a monthly close.
{"start":"2022-10","unit":"$","values":\[30.95,35.17,29.14,37.09,39.61,34.92,36.52,26.19,26.01,30.14,28.15,24.47,21.17,26.09,27.22,26.92,23.6,21.07,20.51,25.23,23.43,26.11,24.87,25.87,29.26,32.61,31.73,30.87,26.89,23.68,25.45,32.42,29.33,22.08,23.55,24.41,22.64,20.93,19.76,16.75,10.83,10.02,11.64,10.99,10.32,8.93],"markers":\[{"month":"2022-11","label":"ChatGPT"},{"month":"2024-12","label":"the fuse ends"}],"endLabel":"$8.93"}
Finding 02
## The two-year fuse.
The obvious story would be a crash the week ChatGPT launched. That is not what happened. Through all of 2023 and 2024, Fiverr's stock held around $30 — the market wasn't yet convinced a chatbot could really replace a freelancer. Then it broke. As 2024's models learned to write, translate, and design at usable quality — and 2025's agents learned to chain those tasks together — the stock halved, and halved again.
| Date | Close | What happened |
| --- | --- | --- |
| Feb 2021 | $269.97 | Pandemic peak — we ignore it |
| Oct 2022 | $30.95 | Month before ChatGPT — our baseline |
| Dec 2024 | $31.73 | Two years later — essentially flat |
| Jul 2026 | $8.93 | Down 71% from baseline — the fuse burns out |
AI didn't kill the gig economy on launch day. It took two years — the time the models needed to get good.
Finding 03
## Gig work was a queue of prompts waiting to happen.
A Fiverr order is, functionally, a prompt: a short instruction — "write me 500 words on X," "translate this," "make a simple logo" — fulfilled by whoever is cheapest. That is the exact interface of a language model, minus the wait, the back-and-forth, and the fee. The high-volume, low-complexity tier of freelancing — the entry-level copy, the basic translation, the template design — is precisely the work models do first and best. Fiverr didn't lose the hard jobs. It lost the easy ones, which were most of the volume.
What it means for you
## If you sell hours for simple tasks, price in the machine.
Fiverr is a warning to anyone whose living is a queue of small, well-specified tasks. Abundance came for the bottom of the ladder first. Three lessons.
**1 · Move up the complexity curve.** The commodity tier — generic copy, basic translation, template work — goes first. Judgment, taste, accountability, and context are what survive.
**2 · Sell outcomes, not tasks.** A task is a prompt, and a model does prompts. An outcome — a campaign that converts, a brand a human trusts — is a relationship, and relationships don't disintermediate.
**3 · Read the two-year fuse.** The market didn't react on launch day; it reacted when the capability matured. Don't confuse "AI hasn't hurt us yet" with "AI won't" — the lag is the trap.
The data
## Every figure, one table.
| Metric | Value |
| --- | --- |
| Bubble peak (Feb 2021, excluded) | $269.97 |
| Post-bubble baseline (Oct 2022) | $30.95 |
| Two years later (Dec 2024) | $31.73 |
| Latest monthly close (Jul 2026) | $8.93 |
| Change since baseline (Oct 2022 → Jul 2026) | −71.1% |
| Change during the flat window (Oct 2022 → Dec 2024) | +2.5% |
| Search-interest change (2022 → 2025 avg) | −73% |
| Upwork (UPWK), peak-to-trough | −84% |
| Months of history measured | 85 (Jul 2019 – Jul 2026) |
| Source | Yahoo Finance monthly close · Wikimedia pageviews |
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
FAQ
## Questions, answered.
**Wasn't this just the 2021 bubble?** We assume so and discard the peak. The 71% we report is measured from the sober post-bubble base of October 2022, with no mania left to deflate. A bubble can't explain a stock that was flat for two years and then fell.
**Why did it fall in 2025–26, not 2023?** Because that is when the capability arrived. Late-2024 models crossed the quality bar for everyday writing, translation, and design; 2025's agents chained those steps together. The market re-priced the labor when the substitute became real, not when it was announced.
**Didn't Fiverr add AI tools?** Yes — like Shutterstock, it shipped AI features and an AI-services category. It didn't stop the slide, because the threat isn't a missing feature; it's that the core unit of work became free.
**Can I reproduce these numbers?** Yes — monthly closing price for FVRR and UPWK from public market data, and Wikimedia's pageviews API for search interest.
## Are you selling tasks — or outcomes AI can't fake?
Fiverr's commodity tier evaporated the moment a model could do it for free. The work that survives is the work AI answers cite and buyers still trust a human to own. Run a free GEO audit and see where you stand.
The defensible work is what AI can't fake at scale: researched, cited, audited — exactly [how HarperFlow produces content](https://www.harperflow.io/features).
[Run the free audit](https://www.harperflow.io/geo-website-audit)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/dead-internet/getty-images-collapse
# The Getty Images collapse.
Getty Images has been the establishment of licensed imagery for thirty years — the wire-photo archive, the red-carpet shots, the corporate stock library. It went public in July 2022, just as machines learned to generate a passable picture of anything for free. Getty's response was the opposite of Shutterstock's: it sued. Then it lost the case *and* 98% of its value.
$26.15→$0.37 share price · Jul 2022 public debut → Jul 2026 · Yahoo Finance, monthly close
49 months as a public company · Jul 2022 → Jul 2026 · monthly closing price · public market data · **every figure recomputable**
−98.6%
from its July 2022 debut
$26.15 → $0.37
−95%
since Oct 2022, its first settled month
$6.80 → $0.37
Nov 2025
it sued the machine — and lost
UK High Court ruled for Stability AI
−96%
Shutterstock, same sector
proof it isn't just a SPAC
**Isn't Getty just a failed SPAC?** It is a de-SPAC — it went public in July 2022 by merging with a shell company, at the peak of a fad that saw hundreds of such listings crater regardless of fundamentals. That is a real confounder, and we flag it. Two things answer it. First, the dated AI events line up: the slide runs alongside Stable Diffusion's launch and Getty's own lawsuit, not random de-SPAC noise. Second, and decisively: Shutterstock — a normal, long-public company with no SPAC — fell 96% over the same span. When the de-SPAC and the blue-chip incumbent both crater, the shared cause isn't the ticker's plumbing. It's the product.
Finding 01
## A stock-photo giant, public at the worst possible moment.
Getty timed its market debut, through a SPAC merger, to July 2022 — the same summer DALL·E 2, Midjourney, and Stable Diffusion turned image generation into a free utility. The debut euphoria briefly valued it near $26 a share; what came after is the chart below. This is the entire life of Getty Images as a public company, one monthly close at a time.
{"start":"2022-07","unit":"$","values":\[26.15,20.25,6.79,6.8,5.87,5.55,6.36,6.42,5.09,6.43,4.67,4.88,4.96,4.4,6.49,4.23,4.92,5.25,4.28,4.72,4.15,3.68,3.59,3.26,3.81,3.8,3.81,4.15,2.92,2.16,2.55,2.14,1.73,1.91,1.78,1.66,1.77,1.84,1.98,1.88,1.54,1.34,1.31,0.78,0.79,0.77,1.04,0.86,0.37],"markers":\[{"month":"2023-01","label":"Sues Stability AI"},{"month":"2025-11","label":"UK ruling — Getty loses"}],"endLabel":"$0.37"}
Finding 02
## The company that sued the machine.
Where Shutterstock licensed its library to OpenAI, Getty went to war. In early 2023 it sued Stability AI — maker of Stable Diffusion — in both the United States and the United Kingdom, alleging the model was trained on millions of Getty images scraped without permission. It was the marquee copyright case of the AI era. It did not save the stock.
| Date | Close | What happened |
| --- | --- | --- |
| Jul 2022 | $26.15 | Public debut via SPAC merger |
| Feb 2023 | $6.42 | Sues Stability AI in the US and UK |
| Dec 2024 | $2.16 | Down \~92% from debut |
| Nov 2025 | $1.54 | UK High Court: Getty's copyright claims fail |
| Jul 2026 | $0.37 | Down 98.6% |
Getty took generative AI to court in two countries — and lost more in the market than it ever sought in damages.
Finding 03
## Litigation is not a moat.
The verdict, when it came in November 2025, was almost beside the point. Getty had abandoned its central copyright claims mid-trial, and the UK High Court ruled that Stable Diffusion was not an "infringing copy" — a landmark loss for rights-holders. But the market had delivered its own verdict years earlier. You cannot sue your way out of a capability shift: even if Getty had won every claim, the machines that make images would still exist, still improve, and still be free. A lawsuit can set the terms of the new world; it cannot restore the old one.
What it means for you
## Two giants, two strategies, one outcome.
Shutterstock licensed the disruptor. Getty sued it. Both stocks fell more than 95%. That is the most important fact across this pair of reports: when a capability commoditizes your core product, neither embracing it nor fighting it in court changes the trajectory much. Three lessons.
**1 · Strategy can't beat a capability shift head-on.** License it or litigate it — if the thing you sell is now free to generate, the business re-rates either way. What changes the outcome is having something that isn't the commodity.
**2 · Legal rights are a floor, not a business.** Owning the copyright to images matters — but it doesn't make people pay for images they can now generate. Rights protect against theft; they don't create demand.
**3 · The market prices the shift before the courts rule on it.** Getty's stock had fallen 90% before the UK judgment landed. The tape leads the docket.
The data
## Every figure, one table.
| Metric | Value |
| --- | --- |
| Public debut close (Jul 2022) | $26.15 |
| First settled month (Oct 2022) | $6.80 |
| Late-2024 close (Dec 2024) | $2.16 |
| Latest monthly close (Jul 2026) | $0.37 |
| Change from debut | −98.6% |
| Change since Oct 2022 | −94.6% |
| Shutterstock (SSTK), same span | −96% |
| Lawsuit filed | US and UK, early 2023 |
| UK ruling (Nov 2025) | Stability AI prevailed |
| Months of history measured | 49 (Jul 2022 – Jul 2026) |
| Source | Yahoo Finance, monthly close |
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
FAQ
## Questions, answered.
**Isn't Getty just a failed SPAC?** It is a de-SPAC, and de-SPACs cratered broadly, so we flag it. But the dated AI events line up — and decisively, Shutterstock, no SPAC and long public, fell 96% over the same span. When both crater, the cause is the product, not the ticker's plumbing.
**What happened with the lawsuit?** Getty sued Stability AI in the US and UK in 2023. In November 2025 the UK High Court ruled for Stability after Getty dropped its main copyright claims; the US case continues. The stock had already fallen about 90% before the ruling.
**How is this different from Shutterstock?** Opposite strategies, same outcome. Shutterstock licensed its library to OpenAI; Getty sued. Both stocks fell more than 95%. Neither embracing nor fighting the capability changed the trajectory much.
**Can I reproduce these numbers?** Yes — monthly closing price for GETY and SSTK from public market data. Pull the tickers and check.
## You can't litigate your way out of a capability shift.
Getty sued the machine and still lost the market. The businesses that survive are the ones AI answers cite and send buyers to. Run a free GEO audit and find out which side of the shift you're on.
Fighting AI in court didn't save Getty; adapting is the only play left — the approach we lay out in [Webflow AI, used well](https://www.harperflow.io/webflow-ai).
[Run the free audit](https://www.harperflow.io/geo-website-audit)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/dead-internet/quora-collapse
# The Quora collapse.
Quora was where the internet went to ask a person — the open-ended, "what's it actually like to…" questions a search box couldn't answer. Then a chatbot started answering exactly those questions, instantly and for free. Quora didn't file for bankruptcy; it's private, so there's no stock to watch. What we can watch is whether people still bother looking it up — and they don't.
43k→24k monthly Wikipedia pageviews for “Quora” — a public proxy for search interest · 2022 avg → 2025 avg
103 months of public interest data · Jan 2018 → Jul 2026 · Wikimedia pageviews API · **every figure recomputable**
−43%
public search interest
2022 → 2025 monthly average
−67%
from its 2018–21 baseline
\~45k → \~15k pageviews / mo
−98.5%
Stack Overflow, the other Q&A site
questions / mo · Part 1 — the pattern repeats
2023
Quora's answer: launch an AI bot
Poe, its own chatbot aggregator
**A word on the metric, up front.** Quora is privately held — no stock, no published revenue, no recomputable traffic feed. So we do not claim to measure its traffic. What we can measure, from a public and reproducible source, is how many people actively seek Quora out: monthly Wikipedia pageviews for the “Quora” article, a clean proxy for search interest and intent. That signal fell 43% from 2022 to 2025, and roughly two-thirds from its 2018–21 baseline. It measures attention, not sessions — but attention is what a destination lives on, and Quora's is draining.
Finding 01
## The place you went to ask humans.
Quora's whole premise was human: real people answering the open-ended questions search engines couldn't — "what's it actually like to…," "why did…," "how do I…." For a decade it was where the internet went to ask a person. The chart below is public interest in that destination, month by month — how many people still look Quora up. The two tall spikes are one-off news months; watch the drift beneath them, and where the line ends.
{"start":"2018-01","values":\[45717,38625,50957,53417,54490,49783,51704,45818,42508,44874,38817,64260,43913,38503,39859,37779,40922,33023,37159,38119,39166,38890,37117,39451,38296,39525,37303,45071,41570,40365,41235,40274,40622,36200,36390,50639,41140,37814,41572,37245,30152,28158,124955,54432,47580,45877,38545,40943,45966,43527,44410,40491,41948,38571,42730,44875,43527,42228,45908,40228,45784,44308,48384,49891,123151,122707,124220,105243,106545,59872,73047,74358,107585,62322,61383,83198,47643,37394,43162,44000,43972,35651,36375,33646,35008,29182,27925,25615,23369,24478,21838,22262,21876,21619,23430,17569,17087,15875,15022,15727,17363,15485,14249],"markers":\[{"month":"2022-11","label":"ChatGPT launch"}],"endLabel":"\~14k"}
Finding 02
## ChatGPT answers the exact question Quora was for.
The questions Quora specialized in — subjective, exploratory, "explain it to me like a person would" — are precisely the questions a language model answers best, and now answers instantly, for free, with no login and no scrolling past twelve SEO answers. Where you once opened Quora and waited for a human, you now type the question into a chatbot and get a synthesized reply. The demand didn't vanish; it moved to the machine.
| Period | Pageviews / mo | Context |
| --- | --- | --- |
| 2018–2021 | \~45,000 | Quora at its steady peak of public interest |
| 2022 avg | \~42,900 | The year before ChatGPT |
| 2025 avg | \~24,500 | Down 43% |
| 2026 recent | \~15,000 | Down about two-thirds from baseline |
Quora was where you asked a human. Now you ask the machine — and never look Quora up again.
Finding 03
## Then Quora bet on the machine too.
Quora's own response tells the story better than any chart. In 2023 it launched Poe — an app for talking to AI chatbots — and poured its energy there. The company that got rich on human answers concluded that the future of answers was a model. Meanwhile its Q&A feed filled with AI-generated responses and SEO chum, and the human signal that made Quora worth visiting thinned out. A destination built on people quietly became a wrapper around the thing replacing them.
What it means for you
## When the answer is free, the destination is optional.
Quora is the softest-measured case in this series — and the most universal. It shows what happens to any site whose value was "come here to get an answer," once the answer is one prompt away. Three lessons.
**1 · Attention is the leading indicator.** You may not have a stock to watch, but you can watch whether people still seek you out. Declining branded search and direct interest is the tell, long before the traffic report confirms it.
**2 · Be a source, not a stop.** If AI answers cite you and send a fraction of readers your way, you survive as a source. If it answers instead of you, you become a place people used to go.
**3 · Don't out-AI your own reason to exist.** Quora's pivot to a chatbot conceded its own premise. Adding AI is fine; replacing the human thing people came for is how a community becomes a ghost town.
The data
## Every figure, one table.
| Metric | Value |
| --- | --- |
| 2018–2021 baseline (monthly pageviews) | \~45,000 |
| 2022 average | 42,867 |
| 2025 average | 24,514 |
| 2026 recent month | \~15,000 |
| Change, 2022 → 2025 avg | −43% |
| Change from 2018–21 baseline | ≈ −67% |
| What it measures | Public search interest (not traffic) |
| Months of history measured | 103 (Jan 2018 – Jul 2026) |
| Source | Wikimedia pageviews API · en.wikipedia “Quora” |
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
FAQ
## Questions, answered.
**Isn't this just Wikipedia pageviews, not Quora's traffic?** Correct — and we say so plainly. Quora is private, with no recomputable traffic feed, so we measure the cleanest public signal there is: how many people look Quora up. It is a proxy for interest and intent, not a session count, and we report it as exactly that.
**Couldn't Quora's actual traffic be fine?** Parts of it may be — propped up by search and by its Poe side-app. But "fine on paper via SEO and an AI product" is not the same as "a place humans choose to go." Interest is what a destination runs on, and interest is down by two-thirds.
**Why compare it to Stack Overflow?** Because they are the two ends of the Q&A web — one technical, one general — and both are here for the same reason. Stack Overflow's human questions fell 98.5%; Quora's public interest fell too. The machine that answers questions is hard on sites built to answer questions.
**Can I reproduce this?** Yes — the Wikimedia pageviews API is public and free. Query monthly pageviews for the “Quora” article on English Wikipedia and average by year.
## When AI answers for you, are you the source — or the stop?
Quora was the place you went for an answer, until the answer came without it. The sites that survive are the ones AI cites and sends people to. Run a free GEO audit and find out whether AI treats you as a source or skips you.
The answer isn't to stop answering questions — it's to answer them in structured, citable formats, [which is what our pipeline publishes](https://www.harperflow.io/features).
[Run the free audit](https://www.harperflow.io/geo-website-audit)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/dead-internet/shutterstock-collapse
# The Shutterstock collapse.
Shutterstock built a $4-billion business renting the world its stock photos — hundreds of millions of images licensed one download at a time. Then machines learned to make any picture free, instantly, from a sentence. Shutterstock's answer was clever: license its own library to OpenAI, and ship a generator of its own. The market re-rated it anyway — from **$121 a share to about five dollars**.
$121.15→$5.40 share price · Oct 2021 peak → Jul 2026 · Yahoo Finance, monthly close
103 months of price history · Jan 2018 → Jul 2026 · monthly closing price · public market data · **every figure recomputable**
−96%
from its 2021 peak
$121.15 (Oct 2021) → $5.40
−39%
in the clean AI window — no bubble, no merger
$50.03 (Oct 2022) → $30.35 (Dec 2024)
−70%
public search interest in Shutterstock
2022 → 2025 avg · Wikimedia pageviews
−98%
Getty, the other stock-photo giant
same forces, same cliff
**Isn't a stock just sentiment — and wasn't 2021 a bubble?** Both fair, so we don't lean on the peak. Two confounders are stated up front and fenced off: (1) like every growth stock, Shutterstock had already halved from its 2021 high before generative AI mattered — so the honest AI-era anchor is October 2022, not the top; (2) the 2025–26 attempted merger with Getty, terminated July 2026, adds deal noise to the last eighteen months. Even inside the clean window — Oct 2022 to Dec 2024, no bubble and no merger — the stock still fell 39% while the S&P made new highs. Getty, facing the identical forces, fell far more.
Finding 01
## A stock-photo empire, re-rated in real time.
Shutterstock spent a decade as the quiet toll booth of the visual web: every blog header, every ad mockup, every slide deck, one licensed download at a time. It rode that toll to a $121.15 peak in October 2021. What followed wasn't a dip — it was the market marking down the value of a library of stock images in the same window that machines learned to generate them on demand. Every point below is a monthly close.
{"start":"2018-01","unit":"$","values":\[44.26,50.25,48.15,42.14,47.39,47.46,46.07,55.04,54.58,40.88,38.21,36.01,40.01,46.33,46.63,40.45,38.06,39.19,38.37,35.11,36.12,40.58,41.58,42.88,43.33,38.54,32.16,38.0,37.92,34.97,54.34,50.32,52.04,65.45,68.76,71.7,64.99,88.23,89.04,87.18,90.75,98.17,108.49,115.26,113.32,121.15,114.01,110.88,96.97,90.53,93.08,75.72,60.2,57.31,56.5,55.41,50.17,50.03,53.82,52.72,75.27,75.22,72.6,67.0,49.77,48.67,51.45,42.11,38.05,40.68,43.91,48.28,46.97,48.76,45.81,42.71,40.63,38.7,44.22,35.88,35.37,32.09,31.68,30.35,29.52,21.49,18.63,15.96,18.45,18.96,19.16,20.93,20.85,25.03,20.82,19.1,19.85,16.8,16.61,16.17,14.89,13.95,5.4],"markers":\[{"month":"2022-08","label":"Stable Diffusion / DALL·E 2"},{"month":"2025-01","label":"Getty merger"},{"month":"2026-07","label":"Merger ends"}],"endLabel":"$5.40"}
Finding 02
## The month the pictures became free.
In the space of one summer, image generation went from research demo to open utility: DALL·E 2 opened up in April 2022, Midjourney in July, and Stable Diffusion — open-source and free to run — in August. A stock photo now had a competitor that cost nothing and waited for no one. Shutterstock read the room and licensed its library to OpenAI that October, then launched its own generator. The stock kept sliding regardless.
| Date | Close | What happened |
| --- | --- | --- |
| Oct 2021 | $121.15 | All-time peak |
| Aug 2022 | \~$50 | Stable Diffusion released — free, open image generation |
| Oct 2022 | $50.03 | Licenses library to OpenAI; launches its own generator |
| Dec 2024 | $30.35 | Down 39% in the clean window — before any merger |
| Jul 2026 | $5.40 | After the Getty merger collapses |
Shutterstock sold the machines their training data — and the machines took the business.
Finding 03
## Stock photography was the perfect thing for AI to eat.
A stock photo is a generic, on-demand, rights-cleared picture of a thing — "smiling team in a meeting," "laptop on a desk by a window." That is precisely what a diffusion model produces for free, in any style, with no licensing form and no wait. The moat was the library; the library was also the perfect training set. When the thing you sell is *an interchangeable image*, and images become infinite, the warehouse stops being an asset. Shutterstock didn't lose to a competitor — it lost to a capability.
What it means for you
## If your value is interchangeable, AI ships it cheaper.
Shutterstock is a warning to every business whose product is a commodity unit of content — a generic image, a boilerplate paragraph, a template layout. Abundance is the enemy of that model. Three lessons.
**1 · Sell the un-generatable.** Commodity assets go first. Specific, sourced, rights-sensitive, or brand-true work — the things a model can't invent without you — is what survives.
**2 · Own the buyer, not just the inventory.** A library is a moat only until the library becomes free to replicate. A direct relationship with the person who needs the work outlasts any warehouse of assets.
**3 · Watch which giant blinks.** When the market marks the incumbent down before revenue cracks — and marks its twin down harder — it is pricing a capability shift, not a bad quarter. Chegg's stock moved before its income statement did; so did Shutterstock's.
The data
## Every figure, one table.
| Metric | Value |
| --- | --- |
| Peak monthly close (Oct 2021) | $121.15 |
| Pre-AI-wave close (Oct 2022) | $50.03 |
| Clean-window close (Dec 2024) | $30.35 |
| Latest monthly close (Jul 2026) | $5.40 |
| Change from peak | −95.5% |
| Change, clean AI window (Oct 2022 → Dec 2024) | −39.3% |
| Search-interest change (2022 → 2025 avg) | −70% |
| Getty (GETY), same period | −98% |
| Months of history measured | 103 (Jan 2018 – Jul 2026) |
| Source | Yahoo Finance monthly close · Wikimedia pageviews |
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
FAQ
## Questions, answered.
**Was it AI, or just the 2021 bubble bursting?** Both are in the tape — which is why we fence them off. The bubble explains the fall from $121 to about $50 by late 2022; it does not explain the further 39% slide through 2024, a stretch when the S&P made fresh highs. That decline lines up with generative image tools going mainstream, not with macro.
**Didn't Shutterstock embrace AI?** Yes — it licensed its library to OpenAI and shipped its own generator, the textbook "if you can't beat them" move. The stock still fell. Owning a slice of the disruptor did not offset the commoditization of the core product.
**What about the Getty merger?** Announced January 2025, the two stock-photo giants tried to combine — and the deal collapsed in July 2026 under UK regulatory review. We anchor the AI analysis to before the announcement so merger arbitrage doesn't contaminate the signal.
**Can I reproduce these numbers?** Yes — monthly closing price from public market data, plus Wikimedia's public pageviews API for search interest. Pull SSTK and GETY and check.
## Is AI your distribution — or your replacement?
Shutterstock owned the library and still got commoditized. The businesses that survive are the ones AI answers cite and send buyers to. Run a free GEO audit and find out which side of that line you're on.
When generation is free, quality and provenance are the moat — the principle behind [using AI on a Webflow site without shipping slop](https://www.harperflow.io/webflow-ai).
[Run the free audit](https://www.harperflow.io/geo-website-audit)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/dead-internet/stack-overflow-collapse
# The Stack Overflow collapse.
For fifteen years, developers brought Stack Overflow their questions — at the peak, more than 200,000 a month. Four years after ChatGPT, they bring **1,624**. This is what it looks like when AI quietly drains the well that helped train it.
109,301→1,624 questions per month · Nov 2022 (ChatGPT launch) → Jul 2026 · Stack Exchange API
127 monthly data points · Jan 2016 → Jul 2026 · new questions by creation date · official Stack Exchange API · **every figure recomputable**
−98.5%
questions per month, since ChatGPT launched
Nov 2022 → Jul 2026
−99%
versus the all-time monthly peak
\~201,000 in early 2017
−72%
the fall in 2025 alone, year on year
2024 → 2025
1,624
questions this past month — a 15-year low
Jul 2026
**Is this just deleted questions?** No. These are new questions by creation date, pulled first-party from the official Stack Exchange API. Counts reflect currently-live questions, so historic months read slightly low if anything — which makes the collapse an understatement, not an artifact. Stack Overflow was already sliding \~8% a year before 2022; the claim here is about the *inflection*, not that AI began the decline.
Finding 01
## It was already fading. Then it fell off a cliff.
From 2016 to 2022, Stack Overflow's monthly question volume drifted down about 8% a year as search shifted and the easy questions ran out — steady, unremarkable. Then ChatGPT shipped in November 2022, and the gentle fade became a freefall. Below is every calendar month since 2016; the inflection needs no interpretation.
{"start":"2016-01","values":\[180949,188273,201426,196583,189397,184350,176772,181880,171588,175327,175236,162143,176882,175189,201429,178426,186420,178092,179767,177677,162280,169747,168716,145248,160465,153073,172902,162401,167454,154548,160015,158417,144482,160400,149719,132182,149549,146335,161061,153484,151308,135852,150983,137084,136867,152615,148224,132514,146610,145034,155917,182854,186411,171850,166001,148282,141744,141770,134911,133875,139839,131643,148721,135844,133754,128943,123950,122051,119751,118833,119064,112091,119258,113763,123366,114078,115959,111306,110778,112666,103581,105934,109301,96275,96725,85596,87182,68540,66410,63610,62725,59993,52718,52356,50319,42336,47604,45952,44730,42416,40037,31961,31457,27800,24749,23178,20823,17938,18198,15517,15093,11405,9662,7759,6628,5814,6040,5342,5415,4219,3715,3168,3239,2633,2260,2023,1624],"markers":\[{"month":"2022-11","label":"ChatGPT launches"}],"endLabel":"1,624"}
Finding 02
## Each year it roughly halved. Then halved again.
An 8%-a-year drift does not compound into a 98% collapse in four years. A step-change does. Year over year, the decline didn't merely continue — it accelerated.
| Year | Questions | Year over year |
| --- | --- | --- |
| 2016 (peak) | 2,183,926 | — |
| 2022 · ChatGPT launches | 1,336,266 | −13% |
| 2023 | 788,512 | −41% |
| 2024 | 398,649 | −49% |
| 2025 | 111,098 | −72% |
| 2026 · annualized | \~32,000 | −71% |
A site that once took two million questions a year now takes fewer than one a minute.
Finding 03
## And then it locked its own doors.
Here is the part that should unsettle anyone who builds on the open web. The archive AI is emptying is also **barricading itself against AI**. In 2024, Stack Overflow restricted its long-standing public data dump — the very corpus that helped train the models now replacing it — moving it behind a login and, briefly, watermarking it. The site is going quiet *and* pulling up the drawbridge at once. The human-authored knowledge that made the last decade of software possible isn't only shrinking; it is being withdrawn.
What it means for you
## The questions didn't vanish. They moved.
They're now typed into ChatGPT, Claude, and Perplexity — which answer from models trained on the very archive going quiet. For anyone who publishes to be found, three things follow.
**1 · Assume discovery runs through AI answers, not blue links.** Write to be quoted by a model, not merely ranked by a crawler.
**2 · Structure beats volume.** Pages an engine can parse cleanly — clear headings, FAQs, tables, cited sources — get pulled into answers; walls of prose don't.
**3 · Measure continuously.** What AI cites shifts week to week. A snapshot is a photograph; you need the film.
The data
## Every figure, one table.
| Metric | Value |
| --- | --- |
| Peak monthly questions (Mar 2017) | 201,429 |
| Questions, Nov 2022 (ChatGPT launch) | 109,301 |
| Questions, Jul 2026 (latest) | 1,624 |
| Change since ChatGPT | −98.5% |
| Change vs all-time peak | −99.2% |
| 2023 / 2024 / 2025 year-over-year | −41% / −49% / −72% |
| Months measured | 127 (Jan 2016 – Jul 2026) |
| Source | Stack Exchange API |
The author
## Written by Hesham Mashhour
Founder, HarperFlow
About Hesham
A University of Cambridge graduate and medical doctor turned creator — his YouTube channel NeuroEverything has grown past 40,000 subscribers breaking down brain science. He later turned automation engineer, building content pipelines for founders and creators who'd rather build than market. He started HarperFlow on one conviction: great ideas deserve exposure, and the people best at building things usually have the least time to promote them. He writes from the workbench — HarperFlow runs its own pipeline, on its own blog, first.
[LinkedIn](https://www.linkedin.com/in/heshammashhour/) [NeuroEverything](https://www.youtube.com/channel/UCpGGGC5gcs2lLpuJSM-_6qQ) [Crunchbase](https://www.crunchbase.com/person/hesham-mashhour) [Book a call](https://cal.com/hesham.us/)
FAQ
## Questions, answered.
**Is Stack Overflow actually dying, or just changing?** By its core metric — people asking questions — it has lost 98.5% of its volume in four years, and traffic and answers track the same curve. Independent analyses (The Pragmatic Engineer; Slashdot) report the same near-total decline.
**Couldn't this be Google updates or spam filters, not AI?** Those pressures ran throughout the gentle 2016–2022 slide. The 40–70%-a-year freefall begins precisely at ChatGPT's launch — a step-change, not a trend.
**Can I reproduce these numbers?** Yes — one query per month against the public Stack Exchange API. That's the whole point.
## Is your content being quoted — or going quiet?
The same shift that emptied Stack Overflow decides whether AI answers cite you. Run a free GEO audit and see where you stand.
The lesson isn't 'don't use AI' — it's to use it the way we describe in [our guide to AI for Webflow content](https://www.harperflow.io/webflow-ai): grounded in sources, not generated by the thousand.
[Run the free audit](https://www.harperflow.io/geo-website-audit)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/compare
# Every AI content tool,compared — *fairly*.
Sourced, dated side-by-sides against the tools you're weighing — AI publishers, SEO optimizers, AI writers, and AI-visibility trackers. Every claim is dated and sourced, and we name where each one beats us.
[Start free today →](https://app.harperflow.io/auth/register) [Browse comparisons ↓](#grid)
10 comparisons · every competitor fact verified **2 July 2026**
Choose a matchup
## Pick a comparison
Filter by category, then dive into the full breakdown.
All 10 AI publishers SEO optimizers AI writers GEO trackers
[vsAI publisher
HarperFlow vs Soro
The cheaper, more established auto-publisher — HarperFlow adds citation rigour and a veto on every post.See the comparison →](https://www.harperflow.io/compare/soro) [vsAI publisher
HarperFlow vs BabyLoveGrowth
The feature-richest rival — backlinks and free tools included. HarperFlow wins on citation and control.See the comparison →](https://www.harperflow.io/compare/babylovegrowth) [vsSEO optimizer
HarperFlow vs Surfer SEO
Surfer optimizes drafts to rank on Google. HarperFlow writes and publishes them — built to be cited by AI.See the comparison →](https://www.harperflow.io/compare/surferseo) [vsSEO optimizer
HarperFlow vs Frase
Frase speeds up SERP research and briefs. HarperFlow runs the whole pipeline, from topic to published article.See the comparison →](https://www.harperflow.io/compare/frase) [vsSEO optimizer
HarperFlow vs Clearscope
Clearscope grades content quality for SEO. HarperFlow publishes it — and engineers it to be quoted by AI.See the comparison →](https://www.harperflow.io/compare/clearscope) [vsAI writer
HarperFlow vs Jasper
Jasper is a versatile AI writing tool. HarperFlow is a service that researches, audits, and publishes for you.See the comparison →](https://www.harperflow.io/compare/jasper) [vsAI writer
HarperFlow vs Copy.ai
Copy.ai generates drafts and GTM copy. HarperFlow grounds every claim, audits it, and publishes to your CMS.See the comparison →](https://www.harperflow.io/compare/copyai) [vsGEO tracker
HarperFlow vs AthenaHQ
AthenaHQ measures your visibility in AI answers. HarperFlow creates the content that earns the citations.See the comparison →](https://www.harperflow.io/compare/athenahq) [vsGEO tracker
HarperFlow vs Profound
Profound is deep AI-answer analytics. HarperFlow fills the content gaps its dashboards surface.See the comparison →](https://www.harperflow.io/compare/profound) [vsGEO tracker
HarperFlow vs Otterly.AI
Otterly tracks your mentions in AI search. Pair it with HarperFlow to actually act on what it finds.See the comparison →](https://www.harperflow.io/compare/otterly)
How we compare
## Honest by default
The same standard we hold our own articles to.
01
### Only verifiable, dated facts
Every competitor detail is sourced from their public pages and stamped with the date we checked it. Spot something stale? Tell us and we'll re-verify.
02
### We name where they win
Each comparison flags the rows where the other tool genuinely beats HarperFlow — on price, backlinks, breadth, or track record. Fair beats flattering.
03
### No fake countdowns or borrowed numbers
HarperFlow is new, and we say so. No invented ratings, no scarcity timers, no stats we can't actually measure.
## Done comparing? See it on your *own* site.
Try HarperFlow with three full articles, free — researched, audited, and published to your CMS.
[Start free today →](https://app.harperflow.io/auth/register) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/compare/athenahq
# HarperFlow vs AthenaHQ
Measure your AI visibility — or *create* what earns it?
HarperFlow and AthenaHQ get mentioned in the same breath, but they do different jobs. AthenaHQ measures how your brand shows up across AI answer engines and tells you where you're missing. HarperFlow writes the audited, citation-ready content that fills those gaps — and publishes it. Here's a side-by-side look at where each one leads.
[Start free today →](https://app.harperflow.io/auth/register) [See the breakdown ↓](#matrix)
Last verified 2 July 2026 · sourced from athenahq.ai · HarperFlow pricing current 27 July 2026
The short answer
AthenaHQ is a strong AI-visibility measurement platform — it monitors your brand across eight answer engines, tracks citations and share of voice, and recommends fixes. HarperFlow does the other half: it researches, writes, independently audits, and publishes the content that earns those citations, with a human veto before anything goes live. They're *complementary* — most teams measure with one and create with the other.
Choose HarperFlow if
You need to **create** citation-ready content — researched, independently audited, and published to your CMS with a veto on every post.
Choose AthenaHQ if
You need to **measure** your AI visibility — monitoring, citation and share-of-voice analytics, and dashboards across many answer engines.
At a glance
## The scorecard
Two halves of the same problem. Where each tool genuinely leads — including the rows where AthenaHQ comes out ahead.
| | HarperFlow | AthenaHQ |
| --- | --- | --- |
| Primary goal | Get cited by AI — and create the content that earns it | Measure how your brand shows up in AI answers |
| Tracks AI visibility | PartialFree GEO audit — a snapshot, not ongoing monitoring | YesMonitors 8 engines incl. ChatGPT, Perplexity, AI Overviews, GeminiAthenaHQ's edge |
| AI-answer & citation analytics | NoNot an analytics suite | YesCitations, share of voice, sentiment, dashboards, Ask Athena copilotAthenaHQ's edge |
| Creates the content to fix gaps | YesResearches, writes & publishes the article | PartialContent agents + Shopify blog publishing — newer, measurement-first |
| Independent six-dimension audit | YesSeparate judge, six dimensions, auto-rewrite below bar | NoNo six-dimension content audit documented |
| Human veto before publish | YesA veto digest each cycle — nothing goes live silently | NoNo human-veto publish workflow documented |
| Native Webflow publishing | NativeMapped CMS fields & a ready-made 33-field structure | NoShopify blog publishing; no native Webflow |
| Publishing cadence | Up to daily (30 articles a month) | On-demand via content agents |
| Automatic backlinks | NoNot offered | NoNot offered |
| Entry price | From $62/month for Webflow and WordPress 5 articles a month | Free tier; paid from $295/mo |
| Track record | New — we won't borrow numbers | Established GEO platform, industry advisors AthenaHQ's edge |
Legend: green = yes · yellow = partial · grey = no. AthenaHQ details verified 2 July 2026 from athenahq.ai. Out of date? [Tell us](https://www.harperflow.io/contact) and we'll re-check and re-date it.
AthenaHQ tells you where you're *missing*. HarperFlow writes the answer that gets you cited.
Measurement shows the gap. Creating the content — sourced, audited, and published — is the part we built for. Most teams need both.
Being fair
## Where AthenaHQ is the better pick
If your job is to measure and diagnose AI visibility, AthenaHQ leads. Three honest reasons to pick it.
01
### Broad answer-engine coverage
AthenaHQ monitors eight named engines — ChatGPT, Perplexity, Google AI Overviews and AI Mode, Gemini, Claude, Copilot and Grok. If you need one dashboard across all of them, that breadth is real. HarperFlow doesn't offer ongoing monitoring.
02
### Deep visibility analytics
It tracks which sources AI cites, your share of voice against competitors, and brand sentiment, in board-ready dashboards. That's a measurement job HarperFlow doesn't do.
03
### An AI copilot for your data
Ask Athena lets you query your visibility data in plain language, with prescriptive recommendations. If you want to interrogate the numbers, that's genuinely useful.
The deeper cut
## Three questions that decide it
### Measure, or make?
AthenaHQ's core job is measurement: it shows where your brand appears across AI answers, which sources get cited, and how you compare to rivals. It's genuinely good at that. HarperFlow's core job is the opposite end of the same pipeline — turning a gap into a finished, sourced, citation-ready article and getting it live. AthenaHQ has recently added content agents and a Shopify publishing path, but measurement is its centre of gravity; content creation, auditing, and multi-CMS publishing is ours.
### Diagnosis, or treatment?
A monitor can tell you that a competitor owns the answer to a question you should own. It won't write the page that changes that. HarperFlow researches the topic, grounds every claim in a named source, runs it through an independent six-dimension audit, and puts it in your veto digest before its publish date, so you can skip it, reschedule it, or publish it right away. Anything you don't veto goes live on schedule. Diagnosis and treatment are different disciplines — and doing both well usually means two tools.
### Use them together
The straight recommendation isn't "switch." It's "pair them." Let AthenaHQ measure where you're invisible in AI answers, then point HarperFlow at those gaps to create and publish the content that earns the citation — with the audit and the veto in between. Then let AthenaHQ close the loop and confirm whether the new content moved your visibility.
### Publishing platforms and billing
HarperFlow publishes to Webflow, WordPress, and Shopify. Webflow and WordPress subscriptions start at $62/month for 5 articles. Shopify starts at $39.90/month for 5 articles, billed through Shopify. Wix is in development. [Explore HarperFlow for Shopify](https://www.harperflow.io/shopify) or [install from the Shopify App Store](https://apps.shopify.com/harperflow).
Pricing, side by side
## What each one costs
Different jobs, different bills. AthenaHQ prices for measurement; HarperFlow prices for creating and publishing audited content. HarperFlow comparison prices shown here are for Webflow and WordPress. Shopify starts at $39.90/month for 5 articles, billed through Shopify, with a 7-day trial.
HarperFlow
$62/mo to start
- One Flex plan — you pick 5–30 articles a month
- 5 articles $62/mo · 10 articles $119/mo
- 30 articles $297/mo (price per article falls as volume rises)
- Agency & Enterprise from $1,020
- Independent audit + human veto on every plan
7-day free trial · 3 full articles · cancel anytime.
AthenaHQ
$295/mo (Starter)
- Free "Essential" tier with starter credits
- Starter from $295/mo, credit-based (\~3,600 credits)
- Enterprise — custom (contact sales)
- Monitors 8 AI answer engines
- Measurement, analytics & recommendations
Verified 2 July 2026 from athenahq.ai.
Questions
## HarperFlow vs AthenaHQ, answered
### Is HarperFlow an AthenaHQ alternative or a complement?
Mostly a complement. AthenaHQ measures your AI visibility; HarperFlow creates the content that improves it. If you specifically need audited, citation-ready articles published to your CMS, HarperFlow is the tool — and many teams run both.
### Does AthenaHQ create content?
AthenaHQ is primarily a measurement and analytics platform. It has added content agents and can publish GEO blogs via a Shopify integration, but writing audited, citation-ready articles for Webflow, WordPress, and Shopify (Wix is in development) — with an independent six-dimension audit and a human veto — is HarperFlow's core job.
### Does HarperFlow track AI visibility like AthenaHQ?
Only lightly. HarperFlow offers a free GEO audit as a snapshot, but it isn't an ongoing monitoring dashboard. For continuous multi-engine tracking with citation and share-of-voice analytics, AthenaHQ is built for that.
### Can I use HarperFlow and AthenaHQ together?
Yes, and that's our recommendation. Measure the gaps with AthenaHQ, create and publish the content that fills them with HarperFlow, then measure again to confirm the lift.
### Which costs more, HarperFlow or AthenaHQ?
They price for different jobs. HarperFlow starts at $62/month for Webflow and WordPress for 5 articles a month for content creation and publishing. AthenaHQ has a free Essential tier and a paid Starter plan from $295/mo for measurement; Enterprise is custom.
Want to *create* the content that gets you cited — not just measure it?
That's the exact job HarperFlow was built for. Try it on your own site with three full articles, free — and keep your AI-visibility tracker for the other half.
[Start free today →](https://app.harperflow.io/auth/register) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Keep comparing
## Related reading
[Comparison
HarperFlow vs Profound
Against another AI-visibility platform — measurement vs. audited, published content. Read →](https://www.harperflow.io/compare/profound) [Comparison
HarperFlow vs Otterly.AI
Against an affordable AI-search monitor — tracking vs. creating what earns the mention. Read →](https://www.harperflow.io/compare/otterly) [Overview
All comparisons
Every AI content tool, compared side by side — the full hub. Read →](https://www.harperflow.io/compare)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/compare/babylovegrowth
# HarperFlow vs BabyLoveGrowth
The feature-rich toolkit vs the one built to be *cited*.
HarperFlow and BabyLoveGrowth both research, write, and publish AI SEO articles to your CMS on autopilot. BabyLoveGrowth bundles more around that — an automatic backlink network and a suite of free tools. HarperFlow goes narrower and deeper: citation rigour, an independent quality audit, and a human veto **before** anything goes live. Here's a side-by-side look.
[Start free today →](https://app.harperflow.io/auth/register) [See the breakdown ↓](#matrix)
Last verified 2 July 2026 · sourced from babylovegrowth.ai · HarperFlow pricing current 27 July 2026
The short answer
BabyLoveGrowth is the feature-richest option here: it bundles automatic backlinks from a 2,500+ site partner network and a large suite of free SEO and GEO tools that HarperFlow doesn't match. HarperFlow is narrower on purpose — it's built for AI *citation*, with claims that trace to named sources, an independent six-dimension quality audit, and a human veto so nothing goes live silently. If your priority is off-page reach and free tooling, BabyLoveGrowth leads; if it's citation rigour and control over what publishes under your name, HarperFlow is the better fit.
Choose HarperFlow if
You want citation rigour, an independent quality check, and a veto on every post before it publishes.
Choose BabyLoveGrowth if
You want automatic backlinks and the widest set of free SEO/GEO tools, and you're happy to publish on autopilot.
At a glance
## The scorecard
Where each tool genuinely leads — including the rows where BabyLoveGrowth comes out ahead.
| | HarperFlow | BabyLoveGrowth |
| --- | --- | --- |
| Primary goal | Get cited by AI answer engines (GEO) | Organic traffic from Google & AI search |
| Built for AI citation | YesFacts trace to named sources; omits rather than invents | PartialArticles ship with citations & internal links; no published standard to omit unverifiable claims |
| Independent quality audit | YesSeparate judge, six dimensions, auto-rewrite below bar | NoQuality handled inside its own pipeline; no independent audit published |
| Human veto before publish | YesA veto digest each cycle — nothing goes live silently | NoGenerates and publishes on autopilot by default |
| Webflow depth | NativeMapped CMS fields & a ready-made 33-field structure | GenericWebflow is one of several CMS integrations |
| Publishing cadence | Up to daily (30 articles a month) | Up to daily (30 articles/mo) |
| Automatic backlinks | NoNot offered | Yes2,500+ site partner network; valued at $800+/moBabyLoveGrowth's edge |
| Free GEO tools | YesFree GEO audit on the site | YesChatGPT Visibility, AI Overview & LLM Citation checkers, GEO audit + moreBabyLoveGrowth's edge |
| Marketing posture | No fake countdowns or borrowed numbers | Anchored $247→$99 pricing; monthly "only N spots left" scarcity |
| Entry price | From $62/month for Webflow and WordPress 5 articles a month | $99/mo (anchored from $247) |
| Public track record | New — no borrowed numbers | More established — public since 2024, with third-party reviews BabyLoveGrowth's edge |
Legend: green = yes · yellow = partial · grey = no. Last verified 2 July 2026 · sourced from babylovegrowth.ai. Out of date? [Tell us](https://www.harperflow.io/contact) and we'll re-check and re-date it.
More features is a real advantage. Being *cited* — and keeping control — is a different one.
BabyLoveGrowth does more around each article. HarperFlow puts its effort into the article itself — grounded enough to be quoted, and yours to veto before it goes live.
Being fair
## Where BabyLoveGrowth is the better pick
We only show things we can verify — theirs included. Three reasons BabyLoveGrowth may suit you more.
01
### Automatic backlinks at scale
BabyLoveGrowth includes automatic backlinks through a 2,500+ site partner network, which it values at $800+/mo. HarperFlow doesn't offer backlinks — if off-page link building is the job, that's a real gap on our side.
02
### A deep suite of free tools
Beyond a single audit, BabyLoveGrowth publishes free checkers — ChatGPT Visibility, AI Overview, LLM Citation, a GEO audit and more. HarperFlow offers one free GEO audit; on free tooling, they're clearly ahead.
03
### More features in the box
Backlinks, a Reddit brand-authority agent, JSON-LD schema markup, technical-SEO checks, and 20+ languages. If you want the widest feature set bundled into one plan, BabyLoveGrowth is the feature-richest option here.
The deeper cut
## Three questions that decide it
### Backlinks and extras, or citation rigour?
BabyLoveGrowth's pitch is breadth: articles plus automatic backlinks from a 2,500+ site partner network, a Reddit agent, schema markup, and technical-SEO checks. That's genuinely more surface area, and for some teams it's the right trade. HarperFlow spends its effort narrowly — on making each claim traceable to a named source, and omitting anything it can't ground — so an AI answer engine can quote you with attribution. If your bet is that AI search rewards verifiable, citable content, depth on the article can matter more than breadth around it.
### Self-scored, or independently audited?
Both tools generate and check their own output. The difference is who does the checking. BabyLoveGrowth handles quality inside its own generation pipeline. HarperFlow runs a separate audit — an independent judge scoring six dimensions, with an automatic rewrite when a draft falls below the bar — so the thing that grades the article isn't the same thing that wrote it. When quality is the product, that separation matters.
### Scarcity marketing, or honest defaults?
BabyLoveGrowth's pricing page anchors an original $247 against a $99 rate and adds a monthly "only N spots left" scarcity line. That's a common, effective tactic — noted here as fact, not judgement. HarperFlow's stance is to skip fake countdowns and borrowed numbers: the price is the price, and because it's new, it doesn't cite a track record it hasn't earned yet. Which approach you trust is a real part of the decision.
### Publishing platforms and billing
HarperFlow publishes to Webflow, WordPress, and Shopify. Webflow and WordPress subscriptions start at $62/month for 5 articles. Shopify starts at $39.90/month for 5 articles, billed through Shopify. Wix is in development. [Explore HarperFlow for Shopify](https://www.harperflow.io/shopify) or [install from the Shopify App Store](https://apps.shopify.com/harperflow).
Pricing, side by side
## What each one costs
BabyLoveGrowth bundles backlinks into a single plan. HarperFlow's tiers buy the audit, the veto, and citation rigour — starting lower. HarperFlow comparison prices shown here are for Webflow and WordPress. Shopify starts at $39.90/month for 5 articles, billed through Shopify, with a 7-day trial.
HarperFlow
$62/mo to start
- One Flex plan — you pick 5–30 articles a month
- 5 articles $62/mo · 10 articles $119/mo
- 30 articles $297/mo (price per article falls as volume rises)
- Agency & Enterprise from $1,020
- Independent audit + human veto on every plan
7-day free trial · 3 full articles · cancel anytime.
BabyLoveGrowth
$99/mo (from $247)
- All-in-One $99/mo (anchored from $247)
- 30 SEO/LLM articles/mo, published automatically
- Automatic backlinks via 2,500+ partner sites
- Free tools: ChatGPT Visibility, AI Overview, LLM Citation + more
- 20+ languages · WordPress, Webflow, Shopify, Wix, API
3-day free trial. Verified 2 July 2026 from babylovegrowth.ai.
Questions
## HarperFlow vs BabyLoveGrowth, answered
### Is HarperFlow a good BabyLoveGrowth alternative?
Yes — especially if you publish on Webflow and want to be cited by AI with control over what goes live. You give up BabyLoveGrowth's backlink network and broad free-tool suite, and gain citation rigour, an independent quality audit, and a human veto. See our dedicated BabyLoveGrowth alternative page for the switcher's view.
### Does BabyLoveGrowth do backlinks and free tools?
Yes. BabyLoveGrowth includes automatic backlinks through a 2,500+ site partner network it values at $800+/mo, and publishes a suite of free tools — a ChatGPT Visibility Checker, AI Overview Checker, LLM Citation Checker, GEO audit and more. HarperFlow doesn't offer backlinks and ships a single free GEO audit.
### Which is cheaper, HarperFlow or BabyLoveGrowth?
HarperFlow starts at $62/month for Webflow and WordPress for 5 articles a month, below BabyLoveGrowth's $99/mo (anchored from $247). BabyLoveGrowth bundles backlinks into that price; HarperFlow includes an independent audit and a human veto at every tier.
### Can I switch from BabyLoveGrowth to HarperFlow?
Yes. Connect your site, HarperFlow learns your brand voice from your existing pages, and you can trial three full articles before committing.
### Does HarperFlow auto-publish like BabyLoveGrowth?
It can keep the same hands-free cadence — but every article appears in your veto digest before its publish date, so you can skip it, reschedule it, or publish it right away. Anything you don't veto goes live on schedule. BabyLoveGrowth publishes on autopilot by default.
Want to be *cited* by AI — and keep control of every post?
That's the exact job HarperFlow was built for. Try it on your own site with three full articles, free.
[Start free today →](https://app.harperflow.io/auth/register) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Keep comparing
## Related reading
[Alternative
BabyLoveGrowth alternative
The switcher's guide — what you keep, and what changes. Read →](https://www.harperflow.io/babylovegrowth-alternative) [Comparison
HarperFlow vs Soro
The cheaper, more established generalist — how it stacks up. Read →](https://www.harperflow.io/compare/soro) [Overview
All comparisons
Every AI content tool, compared side by side — the full hub. Read →](https://www.harperflow.io/compare)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/compare/clearscope
# HarperFlow vs Clearscope
A content *grade*, or an article that gets *cited*?
Clearscope gives your draft a trusted A+-to-F content grade and clean term guidance so it ranks on Google — it's the best-known content-grading tool there is. HarperFlow is a different kind of tool: it researches, writes, fact-checks, audits, and **publishes** finished articles to your site, built to be cited by AI answer engines — and it starts below Clearscope's price. Here's a side-by-side look — including where Clearscope wins.
[Start free today →](https://app.harperflow.io/auth/register) [See the breakdown ↓](#matrix)
Last verified 2 July 2026 · sourced from clearscope.io · HarperFlow pricing current 27 July 2026
The short answer
Clearscope is the best-known content-grading tool — a trusted A+-to-F grade, clean term recommendations, and a tidy Google Docs workflow. But it *grades* drafts you still write and publish yourself; it isn't a publisher, and it doesn't ground claims in citations. HarperFlow researches, writes, independently audits across six dimensions, and publishes finished articles built to be cited by AI — with a human veto — and it starts well below Clearscope's price, so if you want cited output rather than a grade, HarperFlow is the better fit.
Choose HarperFlow if
You want finished, cited, audited articles published for you — for less than Clearscope's entry price.
Choose Clearscope if
You have writers and want the most trusted content grade and term guidance to optimize their work.
At a glance
## The scorecard
Where each tool genuinely leads — including the rows where Clearscope comes out ahead.
| | HarperFlow | Clearscope |
| --- | --- | --- |
| Primary goal | Get cited by AI answer engines (GEO) | Grade & optimize your drafts to rank on Google |
| Built for AI citation | YesFacts trace to named sources; omits rather than invents | PartialTracks AI-search visibility; grading is term-based, not citation-grounded |
| Independent quality audit | YesSeparate judge, six dimensions, auto-rewrite below bar | NoTrusted content grade, but it's a SERP term score — not a separate editorial audit |
| Human veto before publish | YesA veto digest each cycle — nothing goes live silently | N/ADoesn't publish — you write & publish yourself |
| Webflow depth | NativeMapped CMS fields & a ready-made 33-field structure | NoNo native Webflow — Google Docs & WordPress add-ons |
| Publishing cadence | Up to daily (30 articles a month) | Doesn't publishIt grades & optimizes drafts; you publish them |
| Content grading & reporting | DifferentAudits quality across six dimensions, not a SERP grade | Best-in-classTrusted A+–F grade; Content Inventory auditsClearscope's edge |
| Free GEO tools | YesFree GEO audit on the site | NoNo free GEO audit |
| Marketing posture | No fake countdowns or borrowed numbers | Transparent — no contracts, no hidden costs |
| Entry price | From $62/month for Webflow and WordPress 5 articles a month HarperFlow's edge | From $129/mo (Essentials) |
| Public track record | New — no borrowed numbers | Established, trusted by content teams Clearscope's edge |
Legend: green = yes · yellow = partial · grey = no. Competitor details verified 2 July 2026 from clearscope.io. Out of date? [Tell us](https://www.harperflow.io/contact) and we'll re-check and re-date it.
A grade measures the draft. HarperFlow *delivers* the cited article.
Clearscope tells you how good your draft is. HarperFlow writes, audits, and publishes one built to be cited.
Being fair
## Where Clearscope is the better pick
We only show things we can state honestly — theirs included. Three reasons Clearscope may suit you more.
01
### The most trusted content grade
Clearscope's A+-to-F content grade and term recommendations are widely regarded as the best pure content-grading system. If your workflow is human writers optimizing toward a grade, that's its home turf. HarperFlow audits quality across six dimensions, but it isn't a SERP grade.
02
### A clean workflow for content teams
The Google Docs add-on embeds the grade in the doc, Content Inventory audits large libraries for decay, and every plan includes unlimited users. For editorial teams, that's a genuinely tidy setup.
03
### An established, transparent brand
Clearscope is trusted by established content teams and prices transparently — no contracts, no hidden costs. HarperFlow is new; we'd rather earn our reputation than borrow one.
The deeper cut
## Three questions that decide it
### Optimize, or publish?
Clearscope's job is to grade: it reads the pages ranking for your keyword, builds a topic map of the terms that matter, and gives your draft a content grade you optimize toward. It's the clearest, most trusted version of that workflow. But you still write the draft, and you still publish it. HarperFlow does the writing and the publishing — it researches, drafts, fact-checks, audits, formats for your CMS, and posts the finished article — so what you get is a live page, not a grade to act on.
### Ranked, or cited?
Clearscope's grade measures how comprehensively you cover a topic against the SERP, and it has added AI-search visibility tracking. Comprehensiveness helps. But being cited by an AI answer engine is a different target: HarperFlow grounds each claim in a named source, omits what it can't support, and structures every article to be quoted with attribution. One optimizes the draft's coverage; the other engineers the finished piece to be lifted by an assistant.
### A tool, or a service?
Clearscope is a premium tool your writers operate — and at $129/mo to start, it assumes you have writers. HarperFlow is a service that produces the writing: a steady cadence of finished, independently-audited, cited articles, each waiting in a veto digest before it publishes, starting at $62/month for Webflow and WordPress. You can use both — grade a cornerstone page in Clearscope, and let HarperFlow handle the cited, published cadence around it.
### Publishing platforms and billing
HarperFlow publishes to Webflow, WordPress, and Shopify. Webflow and WordPress subscriptions start at $62/month for 5 articles. Shopify starts at $39.90/month for 5 articles, billed through Shopify. Wix is in development. [Explore HarperFlow for Shopify](https://www.harperflow.io/shopify) or [install from the Shopify App Store](https://apps.shopify.com/harperflow).
Pricing, side by side
## What each one costs
HarperFlow starts below Clearscope — and it writes and publishes, not just grades. Clearscope's price buys the most trusted content grade for teams that already have writers. HarperFlow comparison prices shown here are for Webflow and WordPress. Shopify starts at $39.90/month for 5 articles, billed through Shopify, with a 7-day trial.
HarperFlow
$62/mo to start
- One Flex plan — you pick 5–30 articles a month
- 5 articles $62/mo · 10 articles $119/mo
- 30 articles $297/mo (price per article falls as volume rises)
- Agency & Enterprise from $1,020
- Independent audit + human veto on every plan
7-day free trial · 3 full articles · cancel anytime.
Clearscope
$129/mo to start
- Essentials $129/mo · 20 drafts, 50 pages
- Business $399/mo · 300 pages + account manager
- Enterprise · custom, SSO
- Unlimited users, no contracts
- Grades & optimizes — you write & publish
Verified 2 July 2026 from clearscope.io. No free trial listed.
Questions
## HarperFlow vs Clearscope, answered
### Is HarperFlow a good Clearscope alternative?
If your workflow is writers optimizing toward a content grade, Clearscope is the trusted standard. If you want finished, cited articles written, audited, and published for you — for less than Clearscope's entry price — HarperFlow is the better fit.
### Does Clearscope publish content or do GEO?
Clearscope grades and optimizes drafts and tracks AI-search visibility, but it doesn't publish to your CMS, and its grade is term-based rather than citation-grounded. HarperFlow is built for GEO and publishes finished, cited articles to your site.
### Which is cheaper, HarperFlow or Clearscope?
HarperFlow is cheaper: it starts at $62/month for Webflow and WordPress for 5 articles a month versus Clearscope's $129/mo Essentials plan — and HarperFlow writes and publishes rather than only grading.
### Can I use HarperFlow and Clearscope together?
Yes. Grade a cornerstone page in Clearscope, and let HarperFlow handle the cited, published cadence around it.
### Does Clearscope write and publish articles for me?
No — Clearscope grades and guides content you write, with a Google Docs add-on and a Content Inventory audit. HarperFlow researches, writes, audits, and publishes the article itself, with a human veto before it goes live.
Want *cited*, published articles — for less than a grading tool?
That's the exact job HarperFlow was built for. Try it on your own site with three full articles, free.
[Start free today →](https://app.harperflow.io/auth/register) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Keep comparing
## Related reading
[Comparison
HarperFlow vs Surfer SEO
The best-known on-page SEO optimizer, compared side by side. Read →](https://www.harperflow.io/compare/surferseo) [Comparison
HarperFlow vs Frase
SERP research and briefs vs. cited, published articles. Read →](https://www.harperflow.io/compare/frase) [Overview
All comparisons
Every AI content tool, compared side by side — the full hub. Read →](https://www.harperflow.io/compare)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/compare/copyai
# HarperFlow vs Copy.ai
Drafts you finish yourself, or *cited articles* published for you?
HarperFlow and Copy.ai both use AI to make content — but they're built for different jobs. Copy.ai is a versatile writing tool with a genuine free tier: you prompt it, it drafts marketing copy and runs go-to-market workflows, and you edit, fact-check, and publish yourself. HarperFlow researches with citations, audits every article across six dimensions, and **publishes** straight to your CMS — with a human veto before anything goes live. Here's a side-by-side look, including where Copy.ai wins.
[Start free today →](https://app.harperflow.io/auth/register) [See the breakdown ↓](#matrix)
Last verified 2 July 2026 · sourced from copy.ai · HarperFlow pricing current 27 July 2026
The short answer
Copy.ai is a versatile AI writing tool with a genuine free tier and growing go-to-market workflow automation — strong for generating marketing copy and drafts at low or no cost. HarperFlow does a different job: it researches with citations, runs an independent quality audit, and publishes finished articles straight to your CMS, with a human veto before anything goes live. If you want free, flexible drafting, Copy.ai fits; if you want cited, audited articles published for you, HarperFlow is the better fit.
Choose HarperFlow if
You want research-grounded, cited articles that are independently audited and published to your site — not just drafts to edit.
Choose Copy.ai if
You want a free-to-start, versatile writing tool for marketing copy and go-to-market workflows.
At a glance
## The scorecard
Where each tool genuinely leads — including the rows where Copy.ai comes out ahead.
| | HarperFlow | Copy.ai |
| --- | --- | --- |
| Primary goal | Get cited by AI answer engines (GEO) | Generate marketing copy & drafts |
| Built for AI citation | YesFacts trace to named sources; omits rather than invents | NoDrafts only — you edit, fact-check & publish |
| Independent quality audit | YesSeparate judge, six dimensions, auto-rewrite below bar | NoNo independent audit — you review the draft yourself |
| Human veto before publish | YesA veto digest each cycle — nothing goes live silently | NoDoesn't publish — you paste & publish it yourself |
| Webflow depth | NativeMapped CMS fields & a ready-made 33-field structure | NoNo CMS publishing — copy/paste into Webflow yourself |
| Publishing cadence | Up to daily (30 articles a month) | NoDoesn't publish to your site |
| Automatic backlinks | NoNot offered | NoNot offered |
| Free GEO tools | YesFree GEO audit on the site | NoNo GEO/citation tooling |
| Marketing posture | No fake countdowns or borrowed numbers | Established brand; genuine free tier |
| Entry price | From $62/month for Webflow and WordPress 5 articles a month | Free tier; Pro from $49/mo Copy.ai's edge |
| Public track record | New — no borrowed numbers | Established, widely adopted (since 2020) Copy.ai's edge |
Legend: green = yes · grey = no or not applicable. Out of date? [Tell us](https://www.harperflow.io/contact) and we'll re-check and re-date it. Last verified 2 July 2026 · sourced from copy.ai
Generating copy is the easy part. Getting it *cited, audited, and published* is the part we built for.
Copy.ai drafts copy for you to edit. HarperFlow delivers cited articles to your site — and lets you veto any of them first.
Being fair
## Where Copy.ai is the better pick
We only show things we can state honestly — theirs included. Three reasons Copy.ai may suit you more.
01
### A permanent free tier
Copy.ai offers a genuine free plan (2,000 words/mo and 90+ templates) — you can start at $0. HarperFlow's entry is $62/month for Webflow and WordPress, with a 7-day trial and three full articles. If starting free matters most, that's a real edge.
02
### GTM & workflow automation
Copy.ai has moved into go-to-market workflow automation — chaining prompts and steps across sales and marketing tasks. HarperFlow focuses on cited articles for your CMS, not multi-step GTM workflows.
03
### Broad versatility
Copy.ai handles a wide range of short-form copy: emails, social posts, ads, and product descriptions. HarperFlow focuses on long-form, cited articles published to your site.
The deeper cut
## Three questions that decide it
### A tool, or a service?
Copy.ai is a tool you operate: you open it, prompt it, generate copy or run a workflow, then edit and place the result. The output scales with the hours and skill you put in. HarperFlow is closer to a service that runs on its own — it researches the topic, writes the article, checks the facts, audits the quality, formats it for your CMS, and publishes it on a cadence, so the result is a finished post on your site rather than a draft in an editor. You can even use both: draft copy in Copy.ai, and let HarperFlow handle the steady, cited article output.
### Words, or published, cited articles?
Copy.ai hands you words — quick, flexible, and cheap to start. What you do next is yours: verify the claims, add sources, format for your CMS, and hit publish. HarperFlow ships the whole thing: each claim is grounded in a named source, the piece is structured with takeaways and an FAQ block so an assistant can quote it with attribution, and it lands in your CMS ready to go. If your bet is that AI answers, not just links, drive demand, that difference compounds.
### Generic draft, or grounded & audited?
General-purpose AI writers generate confident prose, but they don't research-with-citations or run an independent check on what they produce — accuracy is on you. HarperFlow grounds claims in real sources, omits what it can't support rather than inventing it, and runs a separate six-dimension audit that rewrites anything below the bar before a human veto. That's the gap between a good first draft and a publishable, cited article.
### Publishing platforms and billing
HarperFlow publishes to Webflow, WordPress, and Shopify. Webflow and WordPress subscriptions start at $62/month for 5 articles. Shopify starts at $39.90/month for 5 articles, billed through Shopify. Wix is in development. [Explore HarperFlow for Shopify](https://www.harperflow.io/shopify) or [install from the Shopify App Store](https://apps.shopify.com/harperflow).
Pricing, side by side
## What each one costs
Copy.ai starts free and is cheaper at entry — but it gives you drafts, while HarperFlow's price buys research, an independent audit, a human veto, and publishing. HarperFlow comparison prices shown here are for Webflow and WordPress. Shopify starts at $39.90/month for 5 articles, billed through Shopify, with a 7-day trial.
HarperFlow
$62/mo to start
- One Flex plan — you pick 5–30 articles a month
- 5 articles $62/mo · 10 articles $119/mo
- 30 articles $297/mo (price per article falls as volume rises)
- Agency & Enterprise from $1,020
- Independent audit + human veto on every plan
7-day free trial · 3 full articles · cancel anytime.
Copy.ai
$0/mo (free tier)
- Free $0 · 2,000 words/mo, 90+ templates
- Pro $49/mo monthly · $36/mo billed annually
- Pro adds unlimited words & up to 5 seats
- Team & Enterprise: custom pricing
- Drafts only — you edit, fact-check & publish
Verified 2 July 2026 from copy.ai.
Questions
## HarperFlow vs Copy.ai, answered
### Is HarperFlow a good Copy.ai alternative?
It depends on the job. Copy.ai is a strong, versatile writing tool with a free tier and go-to-market workflow automation. HarperFlow is built to research with citations, audit, and publish articles to your CMS. If you want drafts and copy, Copy.ai fits; if you want cited, published articles, HarperFlow does — and the two can be used together.
### Does Copy.ai publish to my site or do GEO?
No. Copy.ai generates drafts and marketing copy; you edit, fact-check, and publish them yourself. It isn't a GEO/citation tool and doesn't publish to your CMS on a schedule. HarperFlow researches with citations, audits, and publishes to Webflow, WordPress, and Shopify; Wix is in development.
### Which is cheaper, HarperFlow or Copy.ai?
Copy.ai is cheaper at entry — it has a permanent free tier (2,000 words/mo) and Pro at $49/mo ($36 billed annually), versus HarperFlow's $62/month for Webflow and WordPress for 5 articles a month. But they deliver different things: Copy.ai gives you drafts; HarperFlow gives you cited, audited, published articles.
### Can I use Copy.ai for drafts and HarperFlow to publish?
Yes. You can draft copy in Copy.ai and use HarperFlow for research-grounded, cited articles published straight to your site. They complement each other well.
### What does HarperFlow do that Copy.ai doesn't?
HarperFlow grounds every claim in named sources, runs an independent six-dimension quality audit, gives you a human veto before publish, and publishes natively to your CMS. Copy.ai focuses on generating drafts, marketing copy, and go-to-market workflows for you to edit and publish yourself.
Want cited, audited articles *published* to your site — not just drafts?
That's the exact job HarperFlow was built for. Try it on your own site with three full articles, free.
[Start free today →](https://app.harperflow.io/auth/register) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Keep comparing
## Related reading
[Comparison
HarperFlow vs Jasper
Another general AI writer — versatile marketing copy and templates, compared side by side. Read →](https://www.harperflow.io/compare/jasper) [Comparison
HarperFlow vs Soro
Against an AI publisher that auto-posts — citation rigour and a human veto, side by side. Read →](https://www.harperflow.io/compare/soro) [Overview
All comparisons
Every AI content tool, compared side by side — the full hub. Read →](https://www.harperflow.io/compare)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/compare/frase
# HarperFlow vs Frase
Great at *briefs* — but built to be *cited*?
Frase turns a keyword into a SERP-driven brief in under a minute, then scores your draft as you write — and it has grown into a broader platform that can even publish. HarperFlow is narrower and deeper: it researches, writes, **grounds every claim in a citation**, independently audits, and publishes finished articles built to be cited by AI. Here's a side-by-side look — including where Frase wins.
[Start free today →](https://app.harperflow.io/auth/register) [See the breakdown ↓](#matrix)
Last verified 2 July 2026 · sourced from frase.io · HarperFlow pricing current 27 July 2026
The short answer
Frase is a strong content platform: fast SERP-driven briefs, question research, real-time scoring, and — newer — publishing and an agent that can auto-publish some page types. It's built to help you research and optimize for search, but it scores content against the SERP rather than grounding every claim in named citations, and it has no independent multi-dimension audit or veto digest before publish. HarperFlow is narrower and deeper: citation-grounded articles, a separate six-dimension audit, native Webflow depth, and a human veto — so if being *cited* by AI with control matters most, HarperFlow fits better.
Choose HarperFlow if
You want citation-grounded articles, an independent audit, and a veto before anything publishes — with native Webflow depth.
Choose Frase if
You want best-in-class SERP research and briefs, the lowest entry price, and a broad content toolkit.
At a glance
## The scorecard
Where each tool genuinely leads — including the rows where Frase comes out ahead.
| | HarperFlow | Frase |
| --- | --- | --- |
| Primary goal | Get cited by AI answer engines (GEO) | Research, brief & optimize content for search |
| Built for AI citation | YesFacts trace to named sources; omits rather than invents | PartialGEO scores & AI-visibility tracking; no claim-level citation standard |
| Independent quality audit | YesSeparate judge, six dimensions, auto-rewrite below bar | NoReal-time content score vs the SERP — not a separate, multi-dimension audit |
| Human veto before publish | YesA veto digest each cycle — nothing goes live silently | NoAgent can auto-publish; no veto-digest step published |
| Webflow depth | NativeMapped CMS fields & a ready-made 33-field structure | GenericPublishes to Webflow as one of several CMSes — not field-mapped depth |
| Publishing cadence | Up to daily (30 articles a month) | Via agentAgent auto-publishes some page types (all types on Scale) |
| SERP research & briefs | LightResearches to cite & publish, not to hand you a brief | ExcellentTop-20 SERP briefs in \~60s; PAA & question miningFrase's edge |
| Free GEO tools | YesFree GEO audit on the site | PartialFree Frase Answers widget; no free GEO audit |
| Marketing posture | No fake countdowns or borrowed numbers | Standard SaaS pricing; 7-day free trial |
| Entry price | From $62/month for Webflow and WordPress 5 articles a month | From $39/mo (Starter, billed yearly) Frase's edge |
| Public track record | New — no borrowed numbers | Established, widely reviewed Frase's edge |
Legend: green = yes · yellow = partial · grey = no. Competitor details verified 2 July 2026 from frase.io. Out of date? [Tell us](https://www.harperflow.io/contact) and we'll re-check and re-date it.
A great brief is a starting line. Being *cited* is the finish.
Frase helps you research and optimize. HarperFlow grounds every claim, audits it, and publishes — with your veto.
Being fair
## Where Frase is the better pick
We only show things we can state honestly — theirs included. Three reasons Frase may suit you more.
01
### Best-in-class SERP research & briefs
Frase turns top-20 SERP results into a structured brief in under a minute, with People-Also-Ask and question mining from sources writers miss. For research and outlining, it's excellent. HarperFlow researches to cite and publish, not to hand you a brief.
02
### The lowest entry price and a free trial
Frase's Starter plan is $39/mo billed yearly — below HarperFlow's $62 — with a 7-day free trial and no credit card required. If the monthly number decides it, that gap is real.
03
### A broad, established platform
Beyond optimization, Frase now publishes to WordPress, Webflow, Wix and more, with an agent for some content types — plus a long public review history. HarperFlow is new and deliberately narrower.
The deeper cut
## Three questions that decide it
### Optimize, or publish?
Frase is built around a research loop: give it a keyword, it analyzes the top-ranking pages, and it hands you a brief and an optimization score to write against. It's one of the fastest ways to turn a SERP into an outline, and it has since added publishing and an agent that can auto-publish some page types. HarperFlow overlaps least on research and most on the finish: it writes the full article, grounds every claim, runs an independent six-dimension audit with auto-rewrite, maps it into your Webflow CMS, and routes it to a veto digest before publish.
### Ranked, or cited?
Frase scores content for search and tracks AI visibility, which helps you cover a topic thoroughly. HarperFlow is built specifically for citation: claims trace to named sources, unsupported points are dropped rather than invented, and structure is engineered so an AI answer engine can quote you with attribution. Both care about AI search; they aim at different points in it — coverage versus quotability.
### A tool, or a service?
This is the closest comparison of the three, because Frase now spans research, optimization, and some publishing. The dividing line is control and rigour at publish time: HarperFlow's separate quality judge and a veto digest by default mean nothing reaches your site unaudited or unseen, and its Webflow integration maps fields rather than pushing generic HTML. Frase gives you more research surface area; HarperFlow gives you a deeper, audited path to a published, cited post. Many teams could reasonably use both.
### Publishing platforms and billing
HarperFlow publishes to Webflow, WordPress, and Shopify. Webflow and WordPress subscriptions start at $62/month for 5 articles. Shopify starts at $39.90/month for 5 articles, billed through Shopify. Wix is in development. [Explore HarperFlow for Shopify](https://www.harperflow.io/shopify) or [install from the Shopify App Store](https://apps.shopify.com/harperflow).
Pricing, side by side
## What each one costs
Frase's entry is lower and it's a broad platform. HarperFlow's price buys citation grounding, an independent audit, a veto before publish, and native Webflow depth. HarperFlow comparison prices shown here are for Webflow and WordPress. Shopify starts at $39.90/month for 5 articles, billed through Shopify, with a 7-day trial.
HarperFlow
$62/mo to start
- One Flex plan — you pick 5–30 articles a month
- 5 articles $62/mo · 10 articles $119/mo
- 30 articles $297/mo (price per article falls as volume rises)
- Agency & Enterprise from $1,020
- Independent audit + human veto on every plan
7-day free trial · 3 full articles · cancel anytime.
Frase
$39/mo, billed yearly
- Starter $39/mo · 10 articles, 50 audits
- Professional $103/mo · 40 articles + agent
- Scale $239/mo · 100 articles, auto-publish
- Publishes to WordPress, Webflow, Wix + FraseCMS
- 7-day free trial, no credit card
Verified 2 July 2026 from frase.io. Yearly-billing prices; month-to-month is higher.
Questions
## HarperFlow vs Frase, answered
### Is HarperFlow a good Frase alternative?
If you mainly need SERP research and briefs, Frase is hard to beat. If you want citation-grounded articles that are independently audited and published to your site with a veto step, HarperFlow is the better fit. They overlap enough that some teams use both.
### Does Frase publish content or do GEO?
Frase does more than optimize now — it publishes to WordPress, Webflow, Wix and more, and an agent can auto-publish some page types. It scores for search and tracks AI visibility, but it doesn't ground each claim in named citations the way HarperFlow's GEO approach does.
### Which is cheaper, HarperFlow or Frase?
Frase's entry (Starter, $39/mo billed yearly) is lower than HarperFlow's $62/month for Webflow and WordPress for 5 articles a month. HarperFlow includes an independent quality audit and a human veto on every plan.
### Can I use HarperFlow and Frase together?
Yes. Teams use Frase for research and briefs, then let HarperFlow write, cite, audit, and publish the finished articles.
### How is HarperFlow's audit different from Frase's content score?
Frase's score measures topic coverage against the SERP as you write. HarperFlow runs a separate quality judge across six dimensions after writing, and auto-rewrites anything below the bar before it reaches your veto digest.
Want every claim *cited* — and a veto before it publishes?
That's the exact job HarperFlow was built for. Try it on your own site with three full articles, free.
[Start free today →](https://app.harperflow.io/auth/register) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Keep comparing
## Related reading
[Comparison
HarperFlow vs Surfer SEO
The best-known on-page SEO optimizer, compared side by side. Read →](https://www.harperflow.io/compare/surferseo) [Comparison
HarperFlow vs Clearscope
The trusted content-grading tool, side by side. Read →](https://www.harperflow.io/compare/clearscope) [Overview
All comparisons
Every AI content tool, compared side by side — the full hub. Read →](https://www.harperflow.io/compare)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/compare/jasper
# HarperFlow vs Jasper
A writing tool, or a *publishing service* that gets you cited?
HarperFlow and Jasper both use AI to make content — but they're built for different jobs. Jasper is a writing tool: you prompt it, it drafts marketing copy in your brand voice, and you edit, fact-check, and publish it yourself. HarperFlow researches with citations, audits every article across six dimensions, and **publishes** straight to your CMS — with a human veto before anything goes live. Here's a side-by-side look, including where Jasper wins.
[Start free today →](https://app.harperflow.io/auth/register) [See the breakdown ↓](#matrix)
Last verified 2 July 2026 · sourced from jasper.ai · HarperFlow pricing current 27 July 2026
The short answer
Jasper is a mature, versatile AI writing tool — strong for generating marketing copy, campaigns, and drafts in a consistent brand voice across dozens of templates. HarperFlow does a different job: it researches with citations, runs an independent quality audit, and publishes finished articles straight to your CMS, with a human veto before anything goes live. If you want raw drafts to edit and place yourself, Jasper fits; if you want cited, audited articles published for you, HarperFlow is the better fit.
Choose HarperFlow if
You want research-grounded, cited articles that are independently audited and published to your site — not just drafts to edit.
Choose Jasper if
You want a versatile writing assistant for marketing copy and campaigns, with strong brand voice, templates, and team features.
At a glance
## The scorecard
Where each tool genuinely leads — including the rows where Jasper comes out ahead.
| | HarperFlow | Jasper |
| --- | --- | --- |
| Primary goal | Get cited by AI answer engines (GEO) | Generate marketing copy & drafts |
| Built for AI citation | YesFacts trace to named sources; omits rather than invents | NoDrafts only — you edit, fact-check & publish |
| Independent quality audit | YesSeparate judge, six dimensions, auto-rewrite below bar | NoNo independent audit — you review the draft yourself |
| Human veto before publish | YesA veto digest each cycle — nothing goes live silently | NoDoesn't publish — you paste & publish it yourself |
| Webflow depth | NativeMapped CMS fields & a ready-made 33-field structure | NoNo CMS publishing — copy/paste into Webflow yourself |
| Publishing cadence | Up to daily (30 articles a month) | NoDoesn't publish to your site |
| Automatic backlinks | NoNot offered | NoNot offered |
| Free GEO tools | YesFree GEO audit on the site | NoNo GEO/citation tooling |
| Marketing posture | No fake countdowns or borrowed numbers | Established, mainstream SaaS brand |
| Entry price | From $62/month for Webflow and WordPress 5 articles a month | Pro from $59/mo (annual) · $69 monthly |
| Public track record | New — no borrowed numbers | Long-established, widely adopted (since 2021) Jasper's edge |
Legend: green = yes · grey = no or not applicable. Out of date? [Tell us](https://www.harperflow.io/contact) and we'll re-check and re-date it. Last verified 2 July 2026 · sourced from jasper.ai
Generating words is the easy part. Getting them *cited, audited, and published* is the part we built for.
Jasper writes drafts for you to edit. HarperFlow delivers cited articles to your site — and lets you veto any of them first.
Being fair
## Where Jasper is the better pick
We only show things we can state honestly — theirs included. Three reasons Jasper may suit you more.
01
### Versatile marketing content
Jasper covers far more than articles: ads, emails, social posts, product descriptions, and full campaigns. If you need a broad, all-purpose copy assistant, that range is a genuine strength. HarperFlow focuses on cited, long-form articles for your CMS.
02
### Strong brand voice & templates
Jasper's brand-voice profiles and large library of marketing templates make it fast to produce on-brand copy at volume across many formats. That template breadth is one of the reasons teams pick it.
03
### Enterprise features & integrations
Jasper offers team seats, workflows, an API, and integrations aimed at larger marketing organizations. If you need multi-seat collaboration inside a bigger stack, Jasper has more of that surface area than HarperFlow, which is a focused publishing tool.
The deeper cut
## Three questions that decide it
### A tool, or a service?
Jasper is a tool you operate: you open it, prompt it, generate copy, then edit and place the result. The output scales with the hours and skill you put in. HarperFlow is closer to a service that runs on its own — it researches the topic, writes the article, checks the facts, audits the quality, formats it for your CMS, and publishes it on a cadence, so the result is a finished post on your site rather than a draft in an editor. You can even use both: draft campaign copy in Jasper, and let HarperFlow handle the steady, cited article output.
### Words, or published, cited articles?
Jasper hands you words — fluent, on-brand, and fast. What you do next is yours: verify the claims, add sources, format for your CMS, and hit publish. HarperFlow ships the whole thing: each claim is grounded in a named source, the piece is structured with takeaways and an FAQ block so an assistant can quote it with attribution, and it lands in your CMS ready to go. If your bet is that AI answers, not just links, drive demand, that difference compounds.
### Generic draft, or grounded & audited?
General-purpose AI writers generate confident prose, but they don't research-with-citations or run an independent check on what they produce — accuracy is on you. HarperFlow grounds claims in real sources, omits what it can't support rather than inventing it, and runs a separate six-dimension audit that rewrites anything below the bar before a human veto. That's the gap between a good first draft and a publishable, cited article.
### Publishing platforms and billing
HarperFlow publishes to Webflow, WordPress, and Shopify. Webflow and WordPress subscriptions start at $62/month for 5 articles. Shopify starts at $39.90/month for 5 articles, billed through Shopify. Wix is in development. [Explore HarperFlow for Shopify](https://www.harperflow.io/shopify) or [install from the Shopify App Store](https://apps.shopify.com/harperflow).
Pricing, side by side
## What each one costs
Entry pricing is close — but Jasper gives you drafts, while HarperFlow's price buys research, an independent audit, a human veto, and publishing. HarperFlow comparison prices shown here are for Webflow and WordPress. Shopify starts at $39.90/month for 5 articles, billed through Shopify, with a 7-day trial.
HarperFlow
$62/mo to start
- One Flex plan — you pick 5–30 articles a month
- 5 articles $62/mo · 10 articles $119/mo
- 30 articles $297/mo (price per article falls as volume rises)
- Agency & Enterprise from $1,020
- Independent audit + human veto on every plan
7-day free trial · 3 full articles · cancel anytime.
Jasper
$59/mo (Pro, billed annually)
- Pro $59/mo annually · $69/mo monthly (single user)
- Unlimited words, marketing templates, brand voice
- Business (teams): custom pricing
- 7-day free trial
- Drafts only — you edit, fact-check & publish
Verified 2 July 2026 from jasper.ai. Legacy Creator plan (\~$39–49/mo) is no longer sold to new users.
Questions
## HarperFlow vs Jasper, answered
### Is HarperFlow a good Jasper alternative?
It depends on the job. If you want a general-purpose writing assistant for marketing copy, Jasper is excellent. If you specifically want research-grounded, cited articles that are independently audited and published to your CMS, HarperFlow is built for that — they solve different problems, and many teams use both.
### Does Jasper publish to my site or do GEO?
No. Jasper generates drafts and marketing copy; you edit, fact-check, and publish them yourself. It isn't a GEO/citation tool and doesn't publish to your CMS on a schedule. HarperFlow researches with citations, audits, and publishes to Webflow, WordPress, and Shopify; Wix is in development.
### Which is cheaper, HarperFlow or Jasper?
HarperFlow starts at $62/month for Webflow and WordPress for 5 articles a month; Jasper's Pro is $59/mo billed annually ($69 monthly) for a single user. Jasper is cheaper at entry, but they deliver different things — Jasper gives you drafts, HarperFlow gives you cited, published articles.
### Can I use Jasper for drafts and HarperFlow to publish?
Yes. Some teams draft marketing copy in Jasper and use HarperFlow for research-grounded, cited articles published straight to their site. They complement each other well.
### What does HarperFlow do that Jasper doesn't?
HarperFlow grounds every claim in named sources, runs an independent six-dimension quality audit, gives you a human veto before publish, and publishes natively to your CMS. Jasper focuses on generating drafts and marketing copy for you to edit and publish yourself.
Want cited, audited articles *published* to your site — not just drafts?
That's the exact job HarperFlow was built for. Try it on your own site with three full articles, free.
[Start free today →](https://app.harperflow.io/auth/register) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Keep comparing
## Related reading
[Comparison
HarperFlow vs Copy.ai
Another general AI writer — versatile drafts and a free tier, compared side by side. Read →](https://www.harperflow.io/compare/copyai) [Comparison
HarperFlow vs Soro
Against an AI publisher that auto-posts — citation rigour and a human veto, side by side. Read →](https://www.harperflow.io/compare/soro) [Overview
All comparisons
Every AI content tool, compared side by side — the full hub. Read →](https://www.harperflow.io/compare)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/compare/otterly
# HarperFlow vs Otterly.AI
Track your AI visibility — or *create* what earns it?
HarperFlow and Otterly.AI are often compared, but they do different jobs. Otterly.AI monitors where your brand shows up across AI search — ChatGPT, Google AI Overviews, Perplexity and more — and tells you where you're missing. HarperFlow writes the audited, citation-ready content that fills those gaps, and publishes it. Here's a side-by-side look at where each one leads.
[Start free today →](https://app.harperflow.io/auth/register) [See the breakdown ↓](#matrix)
Last verified 2 July 2026 · sourced from otterly.ai · HarperFlow pricing current 27 July 2026
The short answer
Otterly.AI is an affordable AI-search monitoring tool — it tracks your brand mentions and citations across the major answer engines, benchmarks you against competitors, and gives GEO recommendations. HarperFlow does the other half: it researches, writes, independently audits, and publishes the content that earns those citations, with a human veto before anything goes live. They're *complementary* — most teams track with one and create with the other.
Choose HarperFlow if
You need to **create** citation-ready content — researched, independently audited, and published to your CMS with a veto on every post.
Choose Otterly.AI if
You need to **measure** your AI visibility cheaply — daily multi-engine tracking, brand-mention monitoring, and competitor benchmarking.
At a glance
## The scorecard
Two halves of the same problem. Where each tool genuinely leads — including the rows where Otterly.AI comes out ahead.
| | HarperFlow | Otterly.AI |
| --- | --- | --- |
| Primary goal | Get cited by AI — and create the content that earns it | Measure where your brand shows up in AI answers |
| Tracks AI visibility | PartialFree GEO audit — a snapshot, not ongoing monitoring | YesDaily tracking: ChatGPT, Google AI Overviews, Perplexity, CopilotOtterly.AI's edge |
| AI-answer & citation analytics | NoNot an analytics suite | YesPrompt research, brand-mention tracking, Brand Visibility Index, benchmarkingOtterly.AI's edge |
| Creates the content to fix gaps | YesResearches, writes & publishes the article | NoRecommends fixes; doesn't write or publish |
| Independent six-dimension audit | YesSeparate judge, six dimensions, auto-rewrite below bar | NoScores page AI-readiness, not article QA |
| Human veto before publish | YesA veto digest each cycle — nothing goes live silently | NoDoesn't create or publish content |
| Native Webflow publishing | NativeMapped CMS fields & a ready-made 33-field structure | NoDoesn't publish content |
| Publishing cadence | Up to daily (30 articles a month) | Not a publisher — a monitor |
| Automatic backlinks | NoNot offered | NoNot offered |
| Entry price | From $62/month for Webflow and WordPress 5 articles a month | From $29/mo Otterly.AI's edge |
| Track record | New — we won't borrow numbers | Established self-serve AI-search tracker Otterly.AI's edge |
Legend: green = yes · yellow = partial · grey = no. Otterly.AI details verified 2 July 2026 from otterly.ai. Out of date? [Tell us](https://www.harperflow.io/contact) and we'll re-check and re-date it.
Otterly.AI tells you where you're *missing*. HarperFlow writes the answer that gets you cited.
Monitoring shows the gap. Creating the content — sourced, audited, and published — is the part we built for. Most teams need both.
Being fair
## Where Otterly.AI is the better pick
If your job is to measure and track AI visibility on a budget, Otterly.AI leads. Three honest reasons to pick it.
01
### A low-cost entry
Otterly.AI starts at $29/mo, with a free trial that doesn't ask for a card. If you want to start measuring AI-search visibility cheaply, that's the lowest entry point here — below HarperFlow's $62/month for Webflow and WordPress.
02
### Daily multi-engine monitoring
It tracks ChatGPT, Google AI Overviews, Perplexity and Copilot daily (with Gemini and Claude as add-ons), across 50+ countries, with a Brand Visibility Index. HarperFlow doesn't monitor.
03
### Reporting and benchmarking
Competitor benchmarking, prompt research and a Looker Studio export make it easy to put AI-visibility data in front of a team. That's measurement work HarperFlow doesn't do.
The deeper cut
## Three questions that decide it
### Measure, or make?
Otterly.AI's whole job is measurement: it shows where your brand appears across AI answers, which prompts you rank for, and how you compare to rivals. It's focused and affordable at that. HarperFlow's job is the opposite end of the same pipeline — turning a gap into a finished, sourced, citation-ready article and getting it live. Otterly.AI gives you the recommendation; HarperFlow does the writing, the audit, and the publishing.
### Diagnosis, or treatment?
A monitor can tell you that a competitor owns the answer to a question you should own. It won't write the page that changes that. HarperFlow researches the topic, grounds every claim in a named source, runs it through an independent six-dimension audit, and puts it in your veto digest before its publish date, so you can skip it, reschedule it, or publish it right away. Anything you don't veto goes live on schedule. Diagnosis and treatment are different disciplines — and doing both well usually means two tools.
### Use them together
The straight recommendation isn't "switch." It's "pair them." Let Otterly.AI track where you're invisible in AI answers, then point HarperFlow at those gaps to create and publish the content that earns the citation — with the audit and the veto in between. Then let Otterly.AI close the loop and confirm whether the new content moved your visibility.
### Publishing platforms and billing
HarperFlow publishes to Webflow, WordPress, and Shopify. Webflow and WordPress subscriptions start at $62/month for 5 articles. Shopify starts at $39.90/month for 5 articles, billed through Shopify. Wix is in development. [Explore HarperFlow for Shopify](https://www.harperflow.io/shopify) or [install from the Shopify App Store](https://apps.shopify.com/harperflow).
Pricing, side by side
## What each one costs
Different jobs, different bills. Otterly.AI's entry is lower and prices for tracking; HarperFlow prices for creating and publishing audited content. HarperFlow comparison prices shown here are for Webflow and WordPress. Shopify starts at $39.90/month for 5 articles, billed through Shopify, with a 7-day trial.
HarperFlow
$62/mo to start
- One Flex plan — you pick 5–30 articles a month
- 5 articles $62/mo · 10 articles $119/mo
- 30 articles $297/mo (price per article falls as volume rises)
- Agency & Enterprise from $1,020
- Independent audit + human veto on every plan
7-day free trial · 3 full articles · cancel anytime.
Otterly.AI
$29/mo to start
- Lite $29/mo · 15 search prompts
- Standard $189/mo · 100 search prompts
- Premium $489/mo · 400 search prompts
- Enterprise — custom (contact sales)
- Free trial, no card required
Verified 2 July 2026 from otterly.ai.
Questions
## HarperFlow vs Otterly.AI, answered
### Is HarperFlow an Otterly.AI alternative or a complement?
Mostly a complement. Otterly.AI measures your AI visibility; HarperFlow creates the content that improves it. If you specifically need audited, citation-ready articles published to your CMS, HarperFlow is the tool — and many teams run both.
### Does Otterly.AI create content?
No. Otterly.AI monitors and tracks your AI-search visibility and gives GEO recommendations, but it doesn't write or publish articles. Creating audited, citation-ready content and publishing it to your CMS is HarperFlow's job.
### Does HarperFlow track AI visibility like Otterly.AI?
Only lightly. HarperFlow offers a free GEO audit as a snapshot, but it isn't an ongoing monitoring dashboard. For continuous multi-engine tracking, brand-mention monitoring and competitor benchmarking, Otterly.AI is built for that.
### Can I use HarperFlow and Otterly.AI together?
Yes, and that's our recommendation. Track the gaps with Otterly.AI, create and publish the content that fills them with HarperFlow, then track again to confirm the lift.
### Which is cheaper, HarperFlow or Otterly.AI?
Otterly.AI's entry price ($29/mo) is lower than HarperFlow's ($62/month for Webflow and WordPress) — but they do different jobs. Otterly.AI's fee is for tracking; HarperFlow's is for creating and publishing audited content, with an independent audit and a human veto at every tier.
Want to *create* the content that gets you cited — not just track it?
That's the exact job HarperFlow was built for. Try it on your own site with three full articles, free — and keep your AI-search tracker for the other half.
[Start free today →](https://app.harperflow.io/auth/register) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Keep comparing
## Related reading
[Comparison
HarperFlow vs AthenaHQ
Against an AI-visibility platform — measurement vs. audited, published content. Read →](https://www.harperflow.io/compare/athenahq) [Comparison
HarperFlow vs Profound
Against enterprise AI-visibility analytics — tracking vs. creating what earns the mention. Read →](https://www.harperflow.io/compare/profound) [Overview
All comparisons
Every AI content tool, compared side by side — the full hub. Read →](https://www.harperflow.io/compare)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/compare/profound
# HarperFlow vs Profound
Measure your AI visibility — or *create* what earns it?
HarperFlow and Profound both live in AI search, but they do different jobs. Profound measures how your brand shows up across AI answer engines — and even which AI crawlers hit your site — so you can see where you're missing. HarperFlow writes the audited, citation-ready content that fills those gaps, and publishes it. Here's a side-by-side look at where each one leads.
[Start free today →](https://app.harperflow.io/auth/register) [See the breakdown ↓](#matrix)
Last verified 2 July 2026 · sourced from tryprofound.com · HarperFlow pricing current 27 July 2026
The short answer
Profound is a mature AI-visibility analytics platform — it tracks brand mentions and citations across major answer engines, benchmarks your share of voice, and even analyses which AI crawlers visit your site. HarperFlow does the other half: it researches, writes, independently audits, and publishes the content that earns those citations, with a human veto before anything goes live. They're *complementary* — most teams measure with one and create with the other.
Choose HarperFlow if
You need to **create** citation-ready content — researched, independently audited, and published to your CMS with a veto on every post.
Choose Profound if
You need to **measure** your AI visibility — answer-engine insights, citation and share-of-voice analytics, and AI-crawler tracking.
At a glance
## The scorecard
Two halves of the same problem. Where each tool genuinely leads — including the rows where Profound comes out ahead.
| | HarperFlow | Profound |
| --- | --- | --- |
| Primary goal | Get cited by AI — and create the content that earns it | Measure how your brand shows up in AI answers |
| Tracks AI visibility | PartialFree GEO audit — a snapshot, not ongoing monitoring | YesMonitors major answer engines, plus AI-crawler analyticsProfound's edge |
| AI-answer & citation analytics | NoNot an analytics suite | YesAnswer Engine Insights, Conversation Explorer, share of voiceProfound's edge |
| Creates the content to fix gaps | YesResearches, writes & publishes the article | PartialProfound Agents can generate & publish — newer, measurement-first |
| Independent six-dimension audit | YesSeparate judge, six dimensions, auto-rewrite below bar | NoNo six-dimension content audit documented |
| Human veto before publish | YesA veto digest each cycle — nothing goes live silently | NoNo human-veto publish workflow documented |
| Native Webflow publishing | NativeMapped CMS fields & a ready-made 33-field structure | NoAgent publishing to CMS/WordPress; no native Webflow |
| Publishing cadence | Up to daily (30 articles a month) | On-demand via agents |
| Automatic backlinks | NoNot offered | NoNot offered |
| Entry price | From $62/month for Webflow and WordPress 5 articles a month | From $99/mo (billed yearly); Enterprise custom |
| Track record | New — we won't borrow numbers | Established, enterprise-adopted platform Profound's edge |
Legend: green = yes · yellow = partial · grey = no. Profound details verified 2 July 2026 from tryprofound.com. Out of date? [Tell us](https://www.harperflow.io/contact) and we'll re-check and re-date it.
Profound tells you where you're *missing*. HarperFlow writes the answer that gets you cited.
Measurement shows the gap. Creating the content — sourced, audited, and published — is the part we built for. Most teams need both.
Being fair
## Where Profound is the better pick
If your job is to measure and diagnose AI visibility at scale, Profound leads. Three honest reasons to pick it.
01
### Deep multi-engine analytics
Profound's Answer Engine Insights track brand mentions, citations, cited sources and share of voice across major answer engines. As a measurement layer, it's mature. HarperFlow doesn't offer ongoing monitoring.
02
### AI-crawler and agent analytics
Its Agent Analytics reads your server logs to show which AI crawlers hit your site and how often. That server-side view is a genuine differentiator HarperFlow doesn't offer.
03
### Conversation and prompt insight
Conversation Explorer and prompt-volume data help you see the real questions people ask AI engines. For research and strategy, that depth is useful.
The deeper cut
## Three questions that decide it
### Measure, or make?
Profound's core job is measurement: it shows where your brand appears across AI answers, which sources get cited, how you compare to rivals, and even which AI crawlers visit your site. It's genuinely good at that. HarperFlow's core job is the opposite end of the same pipeline — turning a gap into a finished, sourced, citation-ready article and getting it live. Profound has added agents that can generate and publish content, but measurement is its centre of gravity; content creation, auditing, and multi-CMS publishing is ours.
### Diagnosis, or treatment?
An analytics platform can tell you that a competitor owns the answer to a question you should own. It won't write the page that changes that. HarperFlow researches the topic, grounds every claim in a named source, runs it through an independent six-dimension audit, and puts it in your veto digest before its publish date, so you can skip it, reschedule it, or publish it right away. Anything you don't veto goes live on schedule. Diagnosis and treatment are different disciplines — and doing both well usually means two tools.
### Use them together
The straight recommendation isn't "switch." It's "pair them." Let Profound measure where you're invisible in AI answers, then point HarperFlow at those gaps to create and publish the content that earns the citation — with the audit and the veto in between. Then let Profound close the loop and confirm whether the new content moved your visibility.
### Publishing platforms and billing
HarperFlow publishes to Webflow, WordPress, and Shopify. Webflow and WordPress subscriptions start at $62/month for 5 articles. Shopify starts at $39.90/month for 5 articles, billed through Shopify. Wix is in development. [Explore HarperFlow for Shopify](https://www.harperflow.io/shopify) or [install from the Shopify App Store](https://apps.shopify.com/harperflow).
Pricing, side by side
## What each one costs
Different jobs, different bills. Profound prices for measurement; HarperFlow prices for creating and publishing audited content. HarperFlow comparison prices shown here are for Webflow and WordPress. Shopify starts at $39.90/month for 5 articles, billed through Shopify, with a 7-day trial.
HarperFlow
$62/mo to start
- One Flex plan — you pick 5–30 articles a month
- 5 articles $62/mo · 10 articles $119/mo
- 30 articles $297/mo (price per article falls as volume rises)
- Agency & Enterprise from $1,020
- Independent audit + human veto on every plan
7-day free trial · 3 full articles · cancel anytime.
Profound
$99/mo (billed yearly)
- Starter $99/mo billed yearly · ChatGPT tracking, 50 prompts
- Growth $399/mo billed yearly
- Enterprise — custom (contact sales)
- Answer Engine Insights, Conversation Explorer, Agent Analytics
- Measurement-first, with content agents
Verified 2 July 2026 from tryprofound.com. Month-to-month price not published.
Questions
## HarperFlow vs Profound, answered
### Is HarperFlow a Profound alternative or a complement?
Mostly a complement. Profound measures your AI visibility; HarperFlow creates the content that improves it. If you specifically need audited, citation-ready articles published to your CMS, HarperFlow is the tool — and many teams run both.
### Does Profound create content?
Profound is best known for AI-visibility measurement. It has added Profound Agents that can generate and publish content to CMSs like WordPress. HarperFlow's core is the content itself — every article is fact-grounded, independently audited across six dimensions, held for your veto before it publishes, and supported on Webflow, WordPress, and Shopify.
### Does HarperFlow track AI visibility like Profound?
Only lightly. HarperFlow offers a free GEO audit as a snapshot, but it isn't an ongoing analytics platform. For continuous answer-engine insights, citation analytics and AI-crawler tracking, Profound is built for that.
### Can I use HarperFlow and Profound together?
Yes, and that's our recommendation. Measure the gaps with Profound, create and publish the content that fills them with HarperFlow, then measure again to confirm the lift.
### Which costs more, HarperFlow or Profound?
They price for different jobs. HarperFlow starts at $62/month for Webflow and WordPress for 5 articles a month for content creation and publishing. Profound's lowest public plan is $99/mo billed yearly (Starter), with Growth at $399/mo and a custom Enterprise tier.
Want to *create* the content that gets you cited — not just measure it?
That's the exact job HarperFlow was built for. Try it on your own site with three full articles, free — and keep your AI-visibility platform for the other half.
[Start free today →](https://app.harperflow.io/auth/register) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Keep comparing
## Related reading
[Comparison
HarperFlow vs AthenaHQ
Against another AI-visibility platform — measurement vs. audited, published content. Read →](https://www.harperflow.io/compare/athenahq) [Comparison
HarperFlow vs Otterly.AI
Against an affordable AI-search monitor — tracking vs. creating what earns the mention. Read →](https://www.harperflow.io/compare/otterly) [Overview
All comparisons
Every AI content tool, compared side by side — the full hub. Read →](https://www.harperflow.io/compare)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/compare/soro
# HarperFlow vs Soro
Published on autopilot — built to be *cited*?
HarperFlow and Soro both write and publish AI articles to your site without you lifting a finger. The real difference is what happens **before** publish — and whether you're earning AI-search demand with citations and control, or just posting. Here's a side-by-side look.
[Start free today →](https://app.harperflow.io/auth/register) [See the breakdown ↓](#matrix)
Last verified 2 July 2026 · sourced from trysoro.com · HarperFlow pricing current 27 July 2026
The short answer
Soro is the cheaper, more established generalist — 100+ platforms, a long Trustpilot track record, daily auto-publishing. HarperFlow is built specifically for AI *citation*: every article is fact-grounded, independently audited across six dimensions, and scheduled articles publish unless you veto them. If you run your content on Webflow and want to be quoted by ChatGPT and Google AI — with full control — HarperFlow is the better fit.
Choose HarperFlow if
You want citation rigour, an independent quality check, and a veto on every post before it publishes.
Choose Soro if
Lowest entry price and a long public review history — across 100+ platforms — matter most.
At a glance
## The scorecard
Where each tool genuinely leads — including the rows where Soro comes out ahead.
| | HarperFlow | Soro |
| --- | --- | --- |
| Primary goal | Get cited by AI answer engines (GEO) | Traffic from Google & ChatGPT |
| Built for AI citation | YesFacts trace to named sources; omits rather than invents | PartialTargets AI traffic; no citation-grounding standard published |
| Independent quality audit | YesSeparate judge, six dimensions, auto-rewrite below bar | NoSelf-reported "100/100" score |
| Human veto before publish | YesA veto digest each cycle — nothing goes live silently | NoAuto-publishes by default |
| Webflow depth | NativeMapped CMS fields & a ready-made 33-field structure | GenericOne of 100+ integrations |
| Publishing cadence | Up to daily (30 articles a month) | Daily |
| Automatic backlinks | NoNot offered | NoNot offered |
| Free GEO tools | YesFree GEO audit on the site | NoNone visible |
| Marketing posture | No fake countdowns or borrowed numbers | Persistent "50% OFF" countdown |
| Entry price | From $62/month for Webflow and WordPress 5 articles a month | From $39/mo Soro's edge |
| Public track record | New — no borrowed numbers | 4.9/5 Trustpilot, 1,783 reviews Soro's edge |
Legend: green = yes · yellow = partial · grey = no. Competitor details verified 2 July 2026 from trysoro.com. Out of date? [Tell us](https://www.harperflow.io/contact) and we'll re-check and re-date it.
Getting posted is easy. Getting *cited* — and keeping control — is the part we built for.
Soro publishes while you sleep. HarperFlow makes sure that, when you wake up, you'd have hit publish yourself.
Being fair
## Where Soro is the better pick
We only show numbers we can measure honestly — theirs included. Three reasons Soro may suit you more.
01
### A lower entry price
Soro starts at $39/mo against HarperFlow's $62/month for Webflow and WordPress. If the monthly number is the deciding factor, that gap is real.
02
### A long public track record
Soro shows 4.9/5 from 1,783 Trustpilot reviews. HarperFlow is new — we'd rather earn our numbers than borrow anyone's.
03
### Reach across 100+ platforms
HarperFlow publishes to Webflow, WordPress, and Shopify today; Wix is in development. If your site lives elsewhere, Soro's breadth may matter more than our depth.
The deeper cut
## Three questions that decide it
### Cited, or just ranked?
Soro's promise is traffic — "get customers from Google & ChatGPT." That's a ranking goal, and a good one. But AI answer engines don't rank pages; they quote content they can verify and attribute. HarperFlow is built around that distinction: named citations, snippet-sized takeaways, an eight-question FAQ block, and clean semantic structure, so an assistant can lift your answer with your name on it. If your bet is that search is moving from links to answers, that difference compounds.
### Autopilot, or autopilot with a hand on the wheel?
Both tools run on their own. The question is what happens the moment before something reaches the public internet under your brand. Soro auto-publishes by default. HarperFlow routes every scheduled article into a veto digest first — skip it, reschedule it, publish it right away, or let it go at its slot because you already saw it coming. Same hands-free cadence; a very different amount of control.
### Native, or one of a hundred connectors?
Publishing to 100+ platforms is a breadth strategy. Publishing deeply into Webflow, WordPress, and Shopify is a depth strategy. Wix is in development. On Webflow specifically, depth shows up as correctly mapped CMS fields, Reference-field handling, asset uploads to your CDN, and a collection that's ready on day one — the kind of fit a generic adapter rarely gets exactly right.
### Publishing platforms and billing
HarperFlow publishes to Webflow, WordPress, and Shopify. Webflow and WordPress subscriptions start at $62/month for 5 articles. Shopify starts at $39.90/month for 5 articles, billed through Shopify. Wix is in development. [Explore HarperFlow for Shopify](https://www.harperflow.io/shopify) or [install from the Shopify App Store](https://apps.shopify.com/harperflow).
Pricing, side by side
## What each one costs
Soro's entry is lower. HarperFlow's price buys the audit, the veto, and citation rigour on every plan. HarperFlow comparison prices shown here are for Webflow and WordPress. Shopify starts at $39.90/month for 5 articles, billed through Shopify, with a 7-day trial.
HarperFlow
$62/mo to start
- One Flex plan — you pick 5–30 articles a month
- 5 articles $62/mo · 10 articles $119/mo
- 30 articles $297/mo (price per article falls as volume rises)
- Agency & Enterprise from $1,020
- Independent audit + human veto on every plan
7-day free trial · 3 full articles · cancel anytime.
Soro
$39/mo to start
- From $39/mo (persistent discount banner)
- Daily publishing, brand-voice training
- 100+ platform integrations
- 14-day money-back guarantee
- Auto-publishes by default
Verified 2 July 2026 from trysoro.com.
Questions
## HarperFlow vs Soro, answered
### Is HarperFlow a good Soro alternative?
Yes — especially if you publish on Webflow and want to be cited by AI with full control over what goes live. See our dedicated Soro alternative page for the switcher's view.
### Does Soro do GEO (get you cited by AI)?
Soro targets traffic from Google and ChatGPT, but publishes no specific citation-grounding standard. HarperFlow is built around GEO: every claim traces to a source, with FAQ blocks and structure engineered to be quoted.
### Which is cheaper, HarperFlow or Soro?
Soro's entry price ($39/mo) is lower than HarperFlow's ($62/month for Webflow and WordPress). HarperFlow includes an independent quality audit and a human veto on every plan.
### Can I switch from Soro to HarperFlow?
Yes. Connect your site, HarperFlow learns your brand voice from your existing pages, and you can trial three full articles before committing.
### Does HarperFlow auto-publish like Soro?
It can keep the same hands-free cadence — but every article appears in your veto digest before its publish date, so you can skip it, reschedule it, or publish it right away. Anything you don't veto goes live on schedule.
Want to be *cited* by AI — and keep control of every post?
That's the exact job HarperFlow was built for. Try it on your own site with three full articles, free.
[Start free today →](https://app.harperflow.io/auth/register) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Keep comparing
## Related reading
[Alternative
Soro alternative
The switcher's guide — why teams move on, and what to look for. Read →](https://www.harperflow.io/soro-alternative) [Comparison
HarperFlow vs BabyLoveGrowth
Against the feature-richest competitor — backlinks and free tools included. Read →](https://www.harperflow.io/compare/babylovegrowth) [Overview
All comparisons
Every AI content tool, compared side by side — the full hub. Read →](https://www.harperflow.io/compare)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/compare/surferseo
# HarperFlow vs Surfer SEO
One *grades* your draft. One writes, cites, and *publishes* it.
Surfer SEO scores and guides your draft so it ranks on Google — it's one of the best on-page optimizers there is. HarperFlow is a different kind of tool: it researches, writes, fact-checks, audits, and **publishes** finished articles to your site, built to be cited by AI answer engines. Here's a side-by-side look — including where Surfer wins.
[Start free today →](https://app.harperflow.io/auth/register) [See the breakdown ↓](#matrix)
Last verified 2 July 2026 · sourced from surferseo.com · HarperFlow pricing current 27 July 2026
The short answer
Surfer SEO is one of the best on-page optimizers you can buy — a SERP-driven Content Editor, real-time scoring, deep keyword data, and now AI-search tracking. But it *grades* drafts you still write and publish yourself; it isn't an end-to-end publisher, and it doesn't ground claims in citations. HarperFlow researches, writes, independently audits, and publishes finished articles built to be cited by AI — with a human veto before anything goes live.
Choose HarperFlow if
You want finished, cited, independently-audited articles published to your site — not a score you act on yourself.
Choose Surfer SEO if
You write your own content and want the best on-page SEO scoring and SERP data to optimize it.
At a glance
## The scorecard
Where each tool genuinely leads — including the rows where Surfer SEO comes out ahead.
| | HarperFlow | Surfer SEO |
| --- | --- | --- |
| Primary goal | Get cited by AI answer engines (GEO) | Optimize your drafts to rank on Google & AI search |
| Built for AI citation | YesFacts trace to named sources; omits rather than invents | PartialTracks AI-search visibility and optimizes for it; no claim-level citation grounding |
| Independent quality audit | YesSeparate judge, six dimensions, auto-rewrite below bar | NoReal-time SEO score vs the SERP — not a separate, multi-dimension audit |
| Human veto before publish | YesA veto digest each cycle — nothing goes live silently | N/ADoesn't publish — you write & publish yourself |
| Webflow depth | NativeMapped CMS fields & a ready-made 33-field structure | NoNo native Webflow — WordPress, Google Docs, Contentful, Zapier |
| Publishing cadence | Up to daily (30 articles a month) | Doesn't publishIt optimizes drafts; you publish them |
| On-page SEO optimization | GEO-firstBuilt for citation & structure, not SERP term-density scoring | Best-in-classSERP-driven Content Editor, NLP terms, real-time scoreSurfer's edge |
| Free GEO tools | YesFree GEO audit on the site | Some free toolsFree keyword & AI-content tools; no free GEO audit |
| Marketing posture | No fake countdowns or borrowed numbers | Standard SaaS pricing; 7-day money-back |
| Entry price | From $62/month for Webflow and WordPress 5 articles a month | From $49/mo (Discovery, billed annually) Surfer's edge |
| Public track record | New — no borrowed numbers | Long-established, large user base Surfer's edge |
Legend: green = yes · yellow = partial · grey = no. Competitor details verified 2 July 2026 from surferseo.com. Out of date? [Tell us](https://www.harperflow.io/contact) and we'll re-check and re-date it.
A score tells you what to fix. HarperFlow *writes*, cites, and publishes the fix.
Surfer grades the draft. HarperFlow delivers the finished, cited article — and lets you veto it first.
Being fair
## Where Surfer SEO is the better pick
We only show things we can state honestly — theirs included. Three reasons Surfer may suit you more.
01
### Best-in-class on-page optimization
Surfer's SERP-driven Content Editor, NLP term recommendations, and real-time Content Score are among the best tools for making a draft rank on Google. HarperFlow optimizes for citation and structure, not term-density scoring.
02
### Deeper SEO data & a broader toolkit
SERP Analyzer, keyword research, internal linking, cannibalization reports, and AI-prompt tracking across ChatGPT, Perplexity, and Google. If you want a full on-page SEO workbench, Surfer has more surface area.
03
### A lower entry price and a long track record
Surfer's Discovery plan starts at $49/mo billed annually, below HarperFlow's $62, and it's an established tool with a large user base. HarperFlow is new — we'd rather earn our numbers than borrow anyone's.
The deeper cut
## Three questions that decide it
### Optimize, or publish?
Surfer's job is to make a draft better before it goes out: you write (or generate) a piece, Surfer scores it against the pages currently ranking, and you edit toward that score. It's a genuinely good optimizer. But the draft, the publishing, the CMS formatting, and the upkeep are still yours. HarperFlow starts where Surfer stops — it researches the topic, writes the article, checks the facts, audits the quality, formats it for your CMS, and publishes it, so the output is a finished post on your site rather than a score in a dashboard.
### Ranked, or cited?
Surfer optimizes for ranking on Google, and has added tracking for how you appear in ChatGPT, Perplexity, and Google's AI answers. That visibility tracking is useful. But tracking whether you're mentioned is different from engineering content to be quoted: HarperFlow grounds each claim in a named source, omits what it can't support, and structures articles — takeaways, an FAQ block, clean semantics — so an assistant can lift your answer with attribution. If your bet is that AI answers matter, that difference compounds.
### A tool, or a service?
Surfer is a tool you operate: results scale with the hours and skill you put in. HarperFlow is closer to a service that runs on its own — a cadence of finished articles, each independently audited and waiting in a veto digest before its publish date, so you can skip it, reschedule it, or publish it right away. Anything you don't veto goes live on schedule. You can even use both: optimize a key page in Surfer, and let HarperFlow handle the steady, cited output around it.
### Publishing platforms and billing
HarperFlow publishes to Webflow, WordPress, and Shopify. Webflow and WordPress subscriptions start at $62/month for 5 articles. Shopify starts at $39.90/month for 5 articles, billed through Shopify. Wix is in development. [Explore HarperFlow for Shopify](https://www.harperflow.io/shopify) or [install from the Shopify App Store](https://apps.shopify.com/harperflow).
Pricing, side by side
## What each one costs
Surfer's entry is lower — but it's a scoring tool, not a publisher. HarperFlow's price buys research, writing, an independent audit, a human veto, and publishing. HarperFlow comparison prices shown here are for Webflow and WordPress. Shopify starts at $39.90/month for 5 articles, billed through Shopify, with a 7-day trial.
HarperFlow
$62/mo to start
- One Flex plan — you pick 5–30 articles a month
- 5 articles $62/mo · 10 articles $119/mo
- 30 articles $297/mo (price per article falls as volume rises)
- Agency & Enterprise from $1,020
- Independent audit + human veto on every plan
7-day free trial · 3 full articles · cancel anytime.
Surfer SEO
$49/mo, billed annually
- Discovery $49/mo · 120 documents
- Standard $99/mo · 360 docs + ChatGPT tracking
- Pro $182/mo · multi-platform AI tracking
- Peace of Mind $299/mo · unlimited documents
- Optimizes drafts — you write & publish
Verified 2 July 2026 from surferseo.com. No free trial; 7-day money-back. Monthly billing costs more.
Questions
## HarperFlow vs Surfer SEO, answered
### Is HarperFlow a good Surfer SEO alternative?
It depends what you want. If you need an on-page SEO scoring tool for drafts you write yourself, Surfer is excellent. If you want finished, cited articles researched, audited, and published for you, HarperFlow is the better fit — and many teams use both.
### Does Surfer SEO publish content or do GEO?
Surfer optimizes drafts and tracks AI-search visibility, but it doesn't publish to your CMS on autopilot, and it doesn't ground claims in citations. HarperFlow is built for GEO: every claim traces to a source, and finished articles publish to your site.
### Which is cheaper, HarperFlow or Surfer SEO?
Surfer's entry (Discovery, $49/mo billed annually) is lower than HarperFlow's $62/month for Webflow and WordPress for 5 articles a month. But they do different jobs: Surfer scores your draft; HarperFlow writes, audits, and publishes the article.
### Can I use HarperFlow and Surfer SEO together?
Yes. A common setup is to optimize a cornerstone page in Surfer while HarperFlow handles the steady cadence of cited, published articles around it.
### Does Surfer SEO write to native Webflow fields?
Surfer integrates with WordPress, Google Docs, and Contentful, and doesn't offer native Webflow field mapping. HarperFlow maps content into a ready-made 33-field Webflow CMS structure.
Want cited articles *published* — not just a better score?
That's the exact job HarperFlow was built for. Try it on your own site with three full articles, free.
[Start free today →](https://app.harperflow.io/auth/register) [Run a free GEO audit](https://www.harperflow.io/geo-website-audit)
Keep comparing
## Related reading
[Comparison
HarperFlow vs Frase
SERP research and briefs vs. cited, published articles. Read →](https://www.harperflow.io/compare/frase) [Comparison
HarperFlow vs Clearscope
The trusted content-grading tool, side by side. Read →](https://www.harperflow.io/compare/clearscope) [Overview
All comparisons
Every AI content tool, compared side by side — the full hub. Read →](https://www.harperflow.io/compare)
Privacy choices
Your privacy
---
Source: https://www.harperflow.io/docs/accounts-orgs-and-sites
# Accounts, organizations, and sites
How HarperFlow separates your login, your per-site workspace, and the site itself.
September 26, 2026
HarperFlow keeps three things separate: your **account** (how you sign in and pay), an **organization** (a workspace for a single connected site, with its own plan and learned business profile), and the **site** itself, the actual Webflow, WordPress, or Shopify property that receives published articles. Understanding this upfront makes multi-site and team setups much less confusing later.
Shopify subscriptions are approved and billed through Shopify. See [Shopify setup](https://www.harperflow.io/docs/connect-shopify) and [Plans and pricing](https://www.harperflow.io/docs/plans-and-pricing) for the separate Shopify billing flow. Wix support is in development.
## Solo vs. Agency accounts
When you set up your first site, you pick an account type. You can change it later in settings:
- **Solo**: one brand, one or more sites. Each site is its own workspace with its own plan.
- **Agency**: built for managing multiple client brands. Sites you add still get their own workspace and learned profile, but they draw from one consolidated account-level plan instead of a separate subscription per site.
## One workspace per site
In practice, each site you connect lives in its own organization, with its own connected credential, its own learned business profile (positioning, audience, voice, topics), its own destination collection or blog, and its own autopilot schedule. Adding a second website, even on the same platform, creates a brand-new workspace and walks you through setup again for that site rather than mixing two businesses' content into one profile.
## Shared memory across sites
If you manage more than one site, an organization can turn on shared memory so HarperFlow pools what it learns (brand voice, topic coverage, published history) across those sites instead of treating each one as a blank slate. This is optional and off by default. It's most useful for an agency managing several properties for the same client, or a company with more than one related site.
## Switching between workspaces
The site switcher in the dashboard lists every workspace you belong to. Picking one swaps your active workspace and reloads the dashboard into that site's own data: its articles, its Brain, its settings. See [Managing multiple sites](https://www.harperflow.io/docs/managing-multiple-sites) for how to add, switch, and remove workspaces.
## Team access
You can invite teammates into a workspace rather than sharing a login. Invited members get a role that controls what they can see and change, so you don't have to hand out your own credentials just to loop someone in on reviewing drafts.
> **Note:** Adding a site always creates a new, separate workspace; it does not add a second site to your existing one. If you're looking for "connect one more site," use "Add another site" in the site switcher rather than trying to reconnect from inside your current workspace's setup.
## Related
- [The onboarding wizard, step by step](https://www.harperflow.io/docs/onboarding-wizard)
- [Managing multiple sites](https://www.harperflow.io/docs/managing-multiple-sites)
- [A tour of your dashboard](https://www.harperflow.io/docs/dashboard-tour)
---
---
Source: https://www.harperflow.io/docs/agency-billing
# Agency billing
How Agency's consolidated tiers, overage, and per-client allocations work.
September 26, 2026
Agency puts every client workspace you manage on one consolidated bill, instead of a separate Solo subscription per client.
## Getting started
Agency isn't a self-serve signup. Book a call with HarperFlow's founder, who confirms your tier and rate. Once your account is unlocked, the account owner completes a one-click upgrade from the Plan page, which collects a card and activates consolidated billing. Any client workspaces already on their own Solo subscription keep running as-is until you migrate them onto the Agency plan, so you're never billed twice for the same site.
## The three tiers
Each tier is a flat monthly base fee that includes a set number of articles, plus a per-article rate for anything beyond that:
| Tier | Included articles/mo | Overage rate | Monthly base |
| --- | --- | --- | --- |
| 1 | 120 | $8.50/article | $1,020 |
| 2 | 660 | $7.50/article | $4,950 |
| 3 | 1,600 | $6.00/article | $9,600 |
The base is billed upfront on the 1st of the month and covers your included articles.
## Overage
Articles published beyond your tier's included baseline bill at your per-article rate, in arrears, so each 1st-of-month invoice combines *next* month's base with *last* month's overage. As a safeguard, if unbilled overage reaches $500 mid-month, it's invoiced immediately rather than left to accumulate. This doesn't touch your base or reset your included count for the month.
## Per-client allocations
Add client workspaces and set each one a monthly article allocation. This allocation is a **hard cap**: once a client reaches it, that client's autopilot publishing pauses for the rest of the billing period, and you raise their allocation to let it resume. The allocation resets at the start of each billing period. Billing itself stays aggregate across every client on the account, not itemized per client, so raising one client's allocation never creates a separate bill for that client. It just lets them publish more against your shared base and overage.
## Canceling
Only the account owner can cancel Agency billing. You can end it at your next billing date (the default, where your access and every client's continues through the period you already paid for, with no partial refund), or cancel immediately and revert to Solo right away.
> **Note:** Your per-article overage rate can be adjusted for your account within a roughly $5.00–$9.90 range. Changing it recalculates your base (base = included articles × rate), so the included count stays the same unless that's changed too.
## Related
- [Plans and pricing](https://www.harperflow.io/docs/plans-and-pricing)
- [Take a break](https://www.harperflow.io/docs/take-a-break)
- [Team members and roles](https://www.harperflow.io/docs/team-members-and-roles)
---
---
Source: https://www.harperflow.io/docs/article-states
# Article states
The handful of states an article moves through, from a fresh draft to live on your site.
September 26, 2026
Every article HarperFlow writes has a status that tells you exactly where it sits in the pipeline. This article covers what each one means and how an article moves between them.
## Scheduled
This is the normal outcome of writing an article, whether autopilot or an on-demand run produced it. A scheduled article has a real publish time and sits in your veto window until that time arrives.
## Published
Once a scheduled article's publish time arrives, HarperFlow publishes it to your connected site automatically, provided nobody skipped it first. Published articles are live on your site.
## Skipped
If you click **Skip this one** on a scheduled article before its publish time, it moves to skipped and will not go live. Skipping is the entire review action in this pipeline: there's no separate approval step to click through first, just the choice to stop something you don't want.
## Unpublished
An already-published article can be taken back off your site with **Take this down**. This is different from skipping: the article was live and now isn't, but nothing about it is deleted. See [Unpublish and take-down](https://www.harperflow.io/docs/unpublish-and-take-down).
## Pending review
Most accounts see very little of this state today. It shows up when an article predates the current scheduling behavior, when an article has been restored (see below), or when HarperFlow's sourcing checks held a fresh draft back from auto-scheduling; see [Link verification](https://www.harperflow.io/docs/link-verification). A pending-review article shows a quiet "Add to schedule" action rather than the red skip button, since there's nothing urgent to review yet, just a slot to give it.
## Restoring a skipped or unpublished article
Both skipped and taken-down articles can be restored. Restoring returns an article to pending review with its old publish time cleared, so it needs a new slot before it will go anywhere again.
> **Warning:** If you're regularly seeing articles land in pending review instead of scheduled, that's usually a signal worth investigating, thin sourcing is the most common cause, not a step every article is expected to pass through.
## Related
- [Reviewing before publish](https://www.harperflow.io/docs/reviewing-before-publish)
- [Scheduling and spread](https://www.harperflow.io/docs/scheduling-and-spread)
- [Unpublish and take-down](https://www.harperflow.io/docs/unpublish-and-take-down)
- [Link verification](https://www.harperflow.io/docs/link-verification)
---
---
Source: https://www.harperflow.io/docs/article-url-and-post-markdown-fields
# Article URL and Post Markdown fields
Two CMS fields on your collection: the article's live link and a clean Markdown copy, and how to get them on a collection you already have.
September 26, 2026
Alongside the fields covered in [Custom field mapping](https://www.harperflow.io/docs/custom-field-mapping), every HarperFlow-provisioned Webflow collection can carry two more: **Article URL** and **Post Markdown**. Neither is required to publish; a collection without them keeps working exactly as before.
## Article URL
The article's own live URL once it's published, built from your site's domain and the page it publishes to. Bind it in Webflow's Designer to power things a plain CMS reference field can't: a "read more" link on a blog list, a related-articles widget, or anything else that needs to point at the article itself rather than just render its content inline.
## Post Markdown
The full article body as clean Markdown: the same content as your Rich Text body field, but as plain, portable text instead of HTML. It's meant for reuse outside your site, pasting into a LinkedIn post, a newsletter, or anywhere else that wants Markdown or plain text rather than a web page.
## Getting them on a collection you already have
A brand-new collection HarperFlow creates for you gets both fields automatically. For an existing collection, both setup choices get you there: from your Site page, use **Change where articles go** and pick either **Extend one you have** or **Map onto one as-is** — both add any HarperFlow field your collection is missing, these two included, while leaving your existing fields exactly as they are. The real difference between the two is mapping control, not field creation: **Map onto one as-is** is the path that offers **Review mapping**, where you can hand-pick which of your own fields receives each piece of content instead of accepting the automatic match.
## When values actually appear
A new article gets both fields filled in as soon as it's written. An article that was published before the fields existed on your collection gets them backfilled the next time it's republished or its template is restyled, not retroactively the moment you add the fields; see [Choosing a layout](https://www.harperflow.io/docs/choosing-a-layout) for restyling.
> **Note:** Both fields are Webflow-only today. If you publish to WordPress or Shopify, your articles don't carry an equivalent field yet.
## Related
- [Custom field mapping](https://www.harperflow.io/docs/custom-field-mapping)
- [Connect your Webflow site](https://www.harperflow.io/docs/connect-webflow)
- [Choosing a layout](https://www.harperflow.io/docs/choosing-a-layout)
---
---
Source: https://www.harperflow.io/docs/canceling-your-subscription
# Canceling your subscription
Cancel through the correct billing provider and understand what happens to your published content.
September 26, 2026
The cancellation route depends on how your site is billed. Canceling a subscription does not delete articles already published on your website.
## Webflow and WordPress subscriptions
Sign in as the account owner, open HarperFlow's **Plan** page, and choose **Cancel plan**. Complete the confirmation flow and check the effective end date shown. The payment-method portal is separate from the in-app cancellation control.
## Shopify subscriptions
For an App Store subscription, uninstall HarperFlow from your Shopify admin to stop future billing cycles and revoke the app's access to your store. Review charges in **Settings → Billing**. A pending charge from the current billing cycle can still appear on your next Shopify invoice; uninstalling does not automatically refund it.
Shopify billing is not managed in Stripe. If you need help with a charge or want to discuss keeping the app connected, [contact HarperFlow support](https://www.harperflow.io/contact) before making changes. See [Shopify's app-charge guidance](https://help.shopify.com/en/manual/your-account/manage-billing/billing-charges/types-of-charges/third-party-charges/app-charges).
## Existing and scheduled articles
Articles already published stay on your site unless you remove them. A disconnected Shopify app cannot publish new articles. Review scheduled content and your effective subscription end date before canceling. See [unpublishing an article](https://www.harperflow.io/docs/unpublish-and-take-down) if you want to remove an individual post.
## Multiple sites and returning later
Check which site's subscription you are changing. An agency billing arrangement can cover several sites, while a Shopify app subscription belongs to its store. To return, reinstall or reopen HarperFlow and follow the plan offered for that site. A new installation does not guarantee another free trial.
---
---
Source: https://www.harperflow.io/docs/choosing-a-layout
# Choosing a layout
How article styling works today: one clean default, your accent color, and Webflow's options.
September 26, 2026
Every HarperFlow article publishes in one clean, minimal default style. There's no gallery of themes to pick from. What you control is your accent color, and, on Webflow, a couple of ways to go further.
## The default style and your accent color
From **Blog templates** in your dashboard, set an accent color, auto-detected from your site where possible, or picked yourself. A live preview shows a sample article in that color. New articles pick it up automatically.
## Applying it to articles you've already published
The same page has a **Restyle all published articles** action, which re-renders and re-publishes everything you've already published in your current default style and accent. Useful after you change the accent, or just to freshen older articles.
## If you're on Webflow
Webflow has three tiers of styling, from most to least hands-off:
1. **The Connector App:** installed in the Webflow Designer, it applies a real, native Designer layout component (header, hero, table of contents, and more) built the Webflow way. This is the direction HarperFlow's Webflow styling is moving toward, but it isn't available to install yet. It's pending publication in the Webflow Marketplace.
2. **Automatic responsive styling:** with no installation, HarperFlow injects responsive CSS into your site's head code, so things like data tables reflow into stacked, readable cards on mobile instead of overflowing. This needs the "custom code" permission on your Webflow connection, so reconnect Webflow if your connection predates it.
3. **Do it yourself:** HarperFlow creates the CMS fields on your collection. Bind them to your own collection template design in the Webflow Designer however you like.
## If you're on WordPress, Shopify, or Wix
Articles publish straight into your existing theme's post or article template. To restyle them, edit that template in your platform's own editor. HarperFlow keeps publishing into it either way.
> **Note:** If you've seen older material describing named article themes (Guide, Sidebar, Story, and similar), that gallery was retired. Every article now publishes in the single default style described above, with your accent as the one customization point.
## Related
- [Connect your Webflow site](https://www.harperflow.io/docs/connect-webflow)
- [Custom field mapping](https://www.harperflow.io/docs/custom-field-mapping)
---
---
Source: https://www.harperflow.io/docs/connect-shopify
# Connect your Shopify store
Install HarperFlow, connect your store, choose a blog, and approve billing through Shopify.
September 26, 2026
Connect your store by installing [HarperFlow from the Shopify App Store](https://apps.shopify.com/harperflow). The standard installation connects your store without copying an API token.
## Install and connect
1. Open the HarperFlow listing and install the app on a store you administer.
2. Open HarperFlow from **Apps** in Shopify admin, then sign in or link your HarperFlow account.
3. Complete setup so HarperFlow can learn your store, audience, and brand voice.
4. Choose an existing blog or create a dedicated HarperFlow blog for your articles.
5. Choose your article volume and review the Shopify subscription approval screen before accepting it.
## Your blog and article design
Articles publish to your selected Shopify blog and use your theme's article design. Webflow-specific features such as a provisioned 33-field CMS collection do not apply to Shopify.
## Shopify billing
Shopify subscriptions start at **$39.90 per month for 5 articles**, with a **7-day free trial**. Other published options are 10 articles for $74.90, 20 for $134.90, and 30 for $184.90. Charges run through Shopify Billing, separately from the Webflow and WordPress prices. Review the trial and charge details in Shopify before approving a plan.
Change your volume from HarperFlow's **Plan** page and approve the replacement subscription in Shopify. View charges in Shopify admin under **Settings → Billing**. See [plans and pricing](https://www.harperflow.io/docs/plans-and-pricing) and [canceling your subscription](https://www.harperflow.io/docs/canceling-your-subscription).
## Check your first article
After setup, check the article and its scheduled time in HarperFlow. The veto digest lets you skip a scheduled article before it goes live. Once published, verify it in your Shopify blog and on the public storefront.
If installation or linking fails, use [Shopify connection troubleshooting](https://www.harperflow.io/docs/shopify-and-wix-connection-issues) or [contact support](https://www.harperflow.io/contact). A custom-app connection is a support-assisted exception, not required for a normal App Store installation.
[See HarperFlow for Shopify](https://www.harperflow.io/shopify) · [Install HarperFlow](https://apps.shopify.com/harperflow)
---
---
Source: https://www.harperflow.io/docs/connect-webflow
# Connect your Webflow site
How to link a Webflow site to HarperFlow with a site API token and pick a collection.
September 26, 2026
Webflow is HarperFlow's most complete integration. Connecting takes a token and a couple of minutes.
## Prerequisites
You need a Webflow site with the CMS available, and access to that site's settings to generate an API token.
## Connect your site
The default way to connect is a **site API token**, not Webflow's account-wide OAuth login. Webflow's OAuth popup is unreliable in some browsers, so the token form is the primary path.
1. In Webflow, go to **Site settings → Apps & integrations → API access → Generate token**.
2. Turn on these permissions: **Authorized users** (read), then **Sites**, **CMS**, **Assets**, **Pages**, **Custom code**, and **Components** (all read and write).
3. Paste the generated token into HarperFlow's Connect step and submit.
If you'd rather authorize through Webflow's login flow instead, there's a secondary link for that. Just know some browsers block its popup, and the token above always works.
## Choose where articles go
Once connected, HarperFlow asks where articles should land. You get three options:
- **Create a ready-made collection** (recommended): a brand-new collection with every HarperFlow field already set up.
- **Extend a collection you have:** keeps your existing collection and design, adding only the fields it's missing.
- **Map onto a collection as-is:** matches HarperFlow's fields to whatever you already have, skipping anything with no home.
See [Custom field mapping](https://www.harperflow.io/docs/custom-field-mapping) for what each HarperFlow field does.
## Verify your first draft
After you finish setup (or trigger an on-demand run), open Webflow's CMS panel and check your chosen collection for a new item. A brand-new collection's page starts blank until you style it in the Designer. See [Choosing a layout](https://www.harperflow.io/docs/choosing-a-layout) for your options there.
> **Note:** If the token you paste can already see a site that another HarperFlow workspace has already connected, HarperFlow adds you to that existing workspace instead of creating a duplicate one. You'll see a message about being added to an existing setup rather than the usual "connected" screen.
## Related
- [Custom field mapping](https://www.harperflow.io/docs/custom-field-mapping)
- [Choosing a layout](https://www.harperflow.io/docs/choosing-a-layout)
- [Reconnecting and token health](https://www.harperflow.io/docs/reconnecting-and-token-health)
---
---
Source: https://www.harperflow.io/docs/connect-wix
# Wix support: in development
Wix publishing is in development. Webflow, WordPress, and Shopify are available now.
September 26, 2026
**Wix support is in development.** HarperFlow does not currently offer self-serve Wix installation or publishing. Older instructions describing a Wix App Market installation are not a supported public setup flow.
HarperFlow is available for [Webflow](https://www.harperflow.io/docs/connect-webflow), [WordPress](https://www.harperflow.io/docs/connect-wordpress), and [Shopify](https://www.harperflow.io/docs/connect-shopify). For a Shopify store, [install HarperFlow from the Shopify App Store](https://apps.shopify.com/harperflow).
If you run a Wix site, [contact us](https://www.harperflow.io/contact) to discuss availability. No Wix release date is promised on this page.
---
---
Source: https://www.harperflow.io/docs/connect-wordpress
# Connect your WordPress site
How the WordPress connection, SSO key, and optional plugin work in HarperFlow.
September 26, 2026
HarperFlow's WordPress integration publishes straight into your theme. A new post renders natively the moment it's live, with no page-design step.
> **Note:** WordPress isn't currently offered in self-serve setup; it doesn't appear as a pickable option in the platform picker. This guide describes how the connection works. If you'd like to publish to WordPress, contact HarperFlow support.
## Prerequisites
Your site must be served over **HTTPS**, since Application Passwords, the credential type HarperFlow uses, require it. You'll also need an Administrator or Editor account.
## Connect your site
1. In WordPress, go to **Users → Profile → Application Passwords**, name one "HarperFlow," and generate it. Copy the whole password, including the spaces.
2. In HarperFlow's Connect step, enter your **site URL**, your **WordPress username** (your login, not your display name), and the **Application Password** you just generated.
3. Submit to connect.
## One more step: the SSO key
Right after connecting, HarperFlow shows a one-time SSO key. Copy it and paste it into your site at **wp-admin → HarperFlow → SSO key → Save**. This key lets HarperFlow publish fully-styled articles and enables one-click sign-in from your WordPress admin, and it's shown to you only once, so save it before continuing.
## Choose where articles go
Pick a dedicated HarperFlow category (recommended: a clean, filterable home with its own theme-native archive page) or publish into a category you already use.
## The optional plugin
The free [HarperFlow plugin](https://wordpress.org/plugins/harperflow/) adds SEO titles and descriptions (working with Yoast or RankMath where installed) plus styled FAQ and TL;DR sections. Install it straight from your WordPress admin — **Plugins → Add New**, search **HarperFlow**, then **Install** and **Activate** (a manual `.zip` is also linked in the setup wizard). Articles publish either way; installing it is a pure enhancement, not a requirement.
## Verify your first draft
After finishing setup or triggering an on-demand run, check **Posts** in your WordPress admin, filtered to your chosen category, for the new draft or scheduled post.
## Related
- [Reconnecting and token health](https://www.harperflow.io/docs/reconnecting-and-token-health)
- [The onboarding wizard, step by step](https://www.harperflow.io/docs/onboarding-wizard)
---
---
Source: https://www.harperflow.io/docs/contacting-support
# Contacting support
How to reach HarperFlow directly, in-app or before you've even signed up.
September 26, 2026
There's no ticket queue or chatbot standing between you and the team. A real person reads every message.
## The fastest way
Look for **Contact us** inside the app. It shows up on your Profile page, on billing and plan screens, and a few other spots where you might get stuck. Write what's going on, and it goes straight to the HarperFlow team with your account attached, so a reply lands in your inbox rather than a support portal. You'll also get an automatic confirmation email the moment it sends.
## Before you've signed in
The same kind of form is available on the register and sign-in screens, for problems that happen before you have an account to attach a message to. You'll be asked for your email so a reply has somewhere to go, and you won't get the automatic confirmation copy this way, just the reply itself once someone reads it.
## What to include
A faster reply usually comes down to a bit of context: which workspace or site you mean, what you were trying to do, and the exact wording of any error message you saw. If it's about a specific article or run, mention its title or the date. The team can look it up much faster with something concrete to search for.
## Talking it through instead of writing it out
For some things, extending a trial, setting up Agency billing, an Enterprise conversation, the app offers a short call with HarperFlow's founder instead of a written form. It's the same team either way; a call just makes more sense when there's back-and-forth to work through.
> **Note:** Messages are rate-limited to a handful in a short window, so give the team a moment to reply before sending a follow-up. It hasn't been lost.
## Related
- [Your free trial](https://www.harperflow.io/docs/your-free-trial)
- [Agency billing](https://www.harperflow.io/docs/agency-billing)
- [Publish failure error reference](https://www.harperflow.io/docs/publish-failures-error-reference)
---
---
Source: https://www.harperflow.io/docs/custom-field-mapping
# Custom field mapping
How HarperFlow matches its article fields to your CMS, and what's required vs. optional.
September 26, 2026
HarperFlow articles are made of named fields: title, body, summary, meta description, FAQs, and more. Mapping tells HarperFlow which field in your own CMS each one should write into.
## The three ways to set it up
When you confirm your destination during setup, you choose one of:
- **A ready-made collection or blog** (recommended). HarperFlow creates a brand-new destination with every field already in place, so nothing is ever silently dropped.
- **Extend one you have.** Keeps your existing collection, blog, or category, and adds only the fields it's missing.
- **Map onto one as-is.** Adds any HarperFlow field your collection is missing, the same as Extend, but is the path to **Review mapping**, where you hand-pick which of your own fields each one maps to (or skip a field yourself).
## What's required vs. optional
Only three fields are required to publish at all: **title**, **slug** (the URL segment), and the article **body**. Everything else, including summary, meta title and description, featured image and its alt text, key takeaways, FAQs, reading time, content type, topic cluster, the article's live URL, a Markdown copy of the body, and the [SEO and schema fields](https://www.harperflow.io/docs/seo-and-schema-fields), is optional. If there's no matching field for one of these, HarperFlow simply leaves it out rather than failing to publish. See [Article URL and Post Markdown fields](https://www.harperflow.io/docs/article-url-and-post-markdown-fields) for what those two are for.
## How auto-matching works
HarperFlow suggests a match for each field based on its name and type, and tells you how many it matched automatically. You can review and change any row by hand. A couple of rules apply everywhere: one destination field can only receive one HarperFlow field (you'll see which field is "taken" if you try to double up), and the slug is always locked to your CMS's own built-in slug field rather than something you remap.
## Changing it later
Your mapping isn't locked in after setup. From your Site page, use **Change where articles go** to revisit the same three-way choice and remap fields at any time.
> **Note:** If you rename or delete a field in your collection after setup, HarperFlow's mapping can go stale and publishing pauses until it's re-validated. Check **Mapping health** on your Site page: usually a single click on **Re-validate mapping** fixes it, unless the field is genuinely gone, in which case you'll need to remap it.
## Related
- [Connect your Webflow site](https://www.harperflow.io/docs/connect-webflow)
- [Article URL and Post Markdown fields](https://www.harperflow.io/docs/article-url-and-post-markdown-fields)
- [Choosing a layout](https://www.harperflow.io/docs/choosing-a-layout)
---
---
Source: https://www.harperflow.io/docs/dashboard-tour
# A tour of your dashboard
What each tab in the HarperFlow dashboard is for, once setup is done.
September 26, 2026
Once setup finishes, you land in your workspace. Everything in the sidebar is scoped to whichever site is currently active, so switch sites first if you manage more than one (see [Managing multiple sites](https://www.harperflow.io/docs/managing-multiple-sites)). Here's what each tab does.
## Home
Your landing page: a quick read on what's happening with this site right now.
## Articles
Every article HarperFlow has drafted, scheduled, or published for this site, across every stage (drafted, scheduled and awaiting your review, or already live), including the veto window before anything goes live.
## Brain
A read-only view of what HarperFlow has learned about your brand and how your content is growing over time.
## Portfolio
An account-level rollup across every site you manage. This only appears if you're on an Agency account or you manage more than one site. A single-site Solo account won't see it, since there'd be nothing to roll up.
## Blog templates
Controls how published articles look: your accent color, a live preview, and a way to re-style everything you've already published. See [Choosing a layout](https://www.harperflow.io/docs/choosing-a-layout).
## Site
Your connection's health, where articles publish, and reconnecting if something breaks. See [Reconnecting and token health](https://www.harperflow.io/docs/reconnecting-and-token-health) and [Managing multiple sites](https://www.harperflow.io/docs/managing-multiple-sites).
## Members
Invite teammates into this workspace and manage their access.
## Plan (or Agency Billing)
Your subscription. Solo accounts see a per-site Plan page; Agency accounts see one consolidated Agency Billing page covering every client site instead.
## Usage
A read-only ledger of what's used a credit this period, articles and their trigger, plus the meter showing how many you have left and when it resets. See [The Usage tab](https://www.harperflow.io/docs/the-usage-tab).
## Coming soon
Two tabs, Syndicate and Matchmaker, are labeled "upcoming" in the sidebar. They aren't live features yet. Each one explains what's planned and lets you ask to be notified when it ships.
> **Note:** Portfolio only shows up once you actually have more than one site, or an Agency account. If you're set up with a single site on Solo, don't worry that it's missing; that's expected, not broken.
## Related
- [Go-live checklist](https://www.harperflow.io/docs/go-live-checklist)
- [Managing multiple sites](https://www.harperflow.io/docs/managing-multiple-sites)
- [Choosing a layout](https://www.harperflow.io/docs/choosing-a-layout)
---
---
Source: https://www.harperflow.io/docs/editing-your-brand-voice
# Editing your brand voice
How HarperFlow learns your brand's voice during onboarding, and how to change it any time afterward.
September 26, 2026
HarperFlow writes in your brand's voice from the very first article, built from what you told it during onboarding. You're never locked into that original answer.
## What makes up your brand voice
Your profile includes a voice and tone, a brand archetype, your audience, your positioning, and a one-line description of what you do. HarperFlow derives an initial draft of this from your onboarding answers, and its writer references all of it on every article it produces afterward, from headline phrasing to how directly it talks about your product. A profile might describe a direct, no-nonsense tone aimed at technical buyers, or a warmer, more encouraging one aimed at first-time users; articles read noticeably differently depending on which you've set.
## Where you edit it
Your brand voice lives on your project's settings and is reflected back to you on your [Org Brain](https://www.harperflow.io/docs/org-brain-and-shared-memory) page. Editing any field there applies to articles written from that point forward. It doesn't retroactively rewrite anything already published; past articles keep the voice they were originally written in. You don't need to touch every field at once either; changing just the one that's off, your positioning, say, is enough to shift future articles.
## What changes when you edit it
A change here doesn't go through some separate re-training step. The very next article HarperFlow drafts, whether from autopilot or an on-demand run, reads the updated profile directly, so you can see the effect of an edit in the next article that comes out rather than waiting on a longer cycle.
> **Note:** If your organization has shared memory turned on across multiple sites, the shared org-level voice can take precedence over what's saved on an individual project. See [Org Brain and shared memory](https://www.harperflow.io/docs/org-brain-and-shared-memory) for how that resolves.
## Related
- [Org Brain and shared memory](https://www.harperflow.io/docs/org-brain-and-shared-memory)
- [The onboarding wizard, step by step](https://www.harperflow.io/docs/onboarding-wizard)
- [Quality and trust panel](https://www.harperflow.io/docs/quality-and-trust-panel)
---
---
Source: https://www.harperflow.io/docs/faq-schema-and-answers
# FAQ, schema, and direct answers
Why FAQ sections and structured data are two of the most reliable levers for getting content cited by AI search.
September 26, 2026
If you only have time to change two things about how you structure content for AI search, make them these: add a genuine FAQ section, and add structured data behind the page. Neither is complicated, and both map directly onto how retrieval systems actually work. This chapter covers each in practical detail.
## Why FAQs perform well
A well-written FAQ entry is already close to the ideal shape for a retrieval system: a specific question, followed immediately by a self-contained, plainly stated answer. That's exactly the answer-first structure a generative engine wants to lift and quote. According to GEO Compass's analysis of structured data for answer engine and generative engine optimization, FAQ and Article-type structured content shows some of the strongest citation performance of any content format, precisely because there's no ambiguity about what question a given passage answers.
That doesn't mean bolting a generic "Frequently Asked Questions" box onto the bottom of a page for the sake of it. It means identifying the real questions your audience actually asks, the ones that show up in sales calls, support tickets, and search queries, and answering each one directly and completely in two or three sentences.
## What makes a good FAQ entry
Not all FAQ sections are equal. A few habits separate ones that get cited from ones that don't:
- **Ask the question the way a person actually asks it.** "How much does it cost?" retrieves better than "Pricing information."
- **Answer completely in the first sentence.** Save nuance, exceptions, and caveats for a second sentence; don't make the reader wait for the actual answer.
- **Keep each entry independent.** A retrieval system may lift a single question-and-answer pair with no surrounding context, so each one needs to make sense entirely on its own.
- **Use real numbers and specifics where they exist**, rather than vague reassurance. "Cancel anytime, effective at the end of your current billing period" beats "Flexible cancellation."
A few habits work against you just as reliably. Copying the same generic FAQ block onto every page dilutes the signal that any one page answers anything specific. Inventing questions nobody actually asks, purely to pad out a section, produces exactly the thin, unverifiable content retrieval systems are built to deprioritize. And writing an answer as a teaser, "Learn more below", rather than an actual answer defeats the purpose: the entry needs to resolve the question on the spot, not defer it.
## What structured data actually is
**Structured data**, usually written as JSON-LD, is a block of machine-readable markup embedded in a page that explicitly labels what the page (or a piece of it) represents: this is an Organization, this is an Article with this author and publish date, this block is a FAQPage with these specific question-and-answer pairs. It doesn't change what a visitor sees. It gives machines an explicit, unambiguous answer to "what is this content and what does it claim," instead of leaving them to infer it from prose.
The schema types most worth having on a content-heavy site are:
| Schema type | What it labels | Why it matters for GEO |
| --- | --- | --- |
| `Organization` | Who you are as a business | Anchors your identity consistently across every page |
| `Article` | A piece of content, its author, and its publish date | Signals authorship and freshness explicitly |
| `FAQPage` | Specific question-and-answer pairs on a page | Lets an engine extract exact Q&A pairs with no guesswork |
| `BreadcrumbList` | Where a page sits in your site's structure | Helps establish topical context around a page |
## Why schema and prose aren't a tradeoff
It's tempting to treat structured data as a technical afterthought, something a developer bolts on separately from the actual writing. That framing undersells it. The most reliable approach is to write the visible FAQ content first, in genuinely good question-and-answer form, and then mirror that same content in the FAQPage markup rather than writing two disconnected versions. Well-written FAQ prose and well-formed schema should describe the same thing twice: once for the reader, once for the machine.
> **Note:** Structured data alone doesn't guarantee a citation. It removes ambiguity for a system that has already decided your content is worth retrieving in the first place. The underlying content still has to be accurate, specific, and genuinely useful.
> **Tip:** If you're prioritizing, put FAQ and schema effort on the pages closest to a buying decision, pricing, comparisons, and core product questions, before applying it everywhere. Those are the questions people are actually asking an AI right before they choose.
## How HarperFlow applies this
Every HarperFlow article carries its FAQ content in a structure built for this: an aggregate rich-text field for display, plus eight individually mapped question-and-answer fields, so the underlying CMS always has clean, discrete Q&A pairs available rather than a single undifferentiated block of FAQ text.
## Related
- [Structured content principles](https://www.harperflow.io/docs/structured-content-principles)
- [Topical authority and content clusters](https://www.harperflow.io/docs/topical-authority-and-clusters)
- [Quality and trust panel](https://www.harperflow.io/docs/quality-and-trust-panel)
---
---
Source: https://www.harperflow.io/docs/geo-for-small-teams
# GEO for small teams
A realistic, prioritized way for a small team to start on generative engine optimization with limited time and people.
September 26, 2026
Most of what's written about generative engine optimization assumes a content team with the time to research, write, structure, and maintain a large library. Most businesses don't have that team. This chapter is about what to actually do when you're one person, or a handful of people, with GEO as one responsibility among many.
## Why this is genuinely hard at small scale
It's worth naming the real obstacles honestly rather than pretending they're minor, because underestimating them is exactly why most GEO efforts stall.
**It's a new skill set.** GEO is only a few years old as a named discipline. Most marketers know SEO; far fewer have had reason yet to learn how retrieval systems select sources, what structured data to add, or which formats earn citations. That knowledge gap takes real time to close, even for a motivated generalist.
**It demands genuine volume.** Topical authority comes from depth across many connected articles, not a handful of flagship posts, simply because a real subject has many real sub-questions. A small team's available hours are usually the binding constraint here, not ideas.
**It demands consistency.** Freshness itself functions as a trust signal, and a sporadic publishing pattern actively works against you, not just neutrally. Cadence is often the hardest thing for a stretched team to sustain, precisely because content work is the first thing to slip when something else becomes urgent.
**It demands quality at that volume.** This is the trap. Rushing toward volume with thin, generic, or unchecked content backfires, because verifiability is the entire mechanism GEO depends on. Cutting the quality corner to hit a publishing quota doesn't just fail to help; it can actively undercut the credibility you're trying to build.
Put together, these four demands are exactly why most teams know they should be doing this, intend to start, and never sustain it. The bottleneck is rarely strategy. It's execution, at quality, at volume, indefinitely.
## Where a small team should actually start
Given limited hours, the highest-leverage order is different from a comprehensive checklist built for a large team. Roughly in priority order:
1. **Audit first, for free.** Before spending any effort, ask ChatGPT, Perplexity, and Google's AI answers your two or three most important category questions. This costs nothing but a few minutes and tells you whether you're starting from zero or from something.
2. **Fix your best existing pages before writing new ones.** Take your three or four highest-traffic or highest-intent pages and restructure them to lead with a direct answer, add a genuine FAQ section, and back key claims with a real source. This is far cheaper than producing new content and often has the fastest payoff.
3. **Pick one narrow subject to go deep on, not ten shallow ones.** A small team covering one real subject thoroughly builds more topical authority than the same hours spread across unrelated topics.
4. **Protect a minimal, sustainable cadence over an ambitious one.** One well-structured article every two weeks, kept up for a year, beats an ambitious weekly plan that collapses after six weeks. Consistency is doing most of the work here, not sheer volume.
5. **Add structured data once the content itself is solid.** Schema markup amplifies content that's already good; it doesn't fix content that isn't.
> **Note:** Notice what's not on this list: a large upfront content calendar, a full topic-cluster map, or a dozen new articles in month one. Those are worth building eventually. They're not where a resource-constrained team should spend its first weeks.
## Be honest with yourself about the tradeoff
At some point, a small team has to answer a plain question: can we actually sustain this ourselves, at the quality bar it needs, indefinitely, alongside everything else on our plate? There's no wrong answer, but there is a dishonest one, which is assuming the answer is yes without checking your own track record on similar commitments.
The realistic options are roughly the same ones that existed before GEO had a name: do it yourself with disciplined, protected time; hire into it, whether a freelancer, a part-time specialist, or an agency; or automate the pipeline with a tool built for exactly this problem. None of these is inherently the right answer for every team. What matters is picking deliberately, based on your actual capacity, rather than defaulting into "we'll get to it" and finding, a year later, that no one did.
> **Tip:** If you've tried and failed to sustain a "we'll write two posts a week" plan before, on any topic, take that as real evidence about your team's actual capacity, not as a reason to feel bad. Plan your GEO approach around the capacity you've demonstrated, not the capacity you'd like to have.
## Where automation fits honestly
Tools that automate discovery, drafting, and publishing, HarperFlow among them, exist specifically because the volume-plus-consistency-plus-quality combination above is genuinely difficult to sustain by hand. That's a legitimate path for a small team, provided whatever you use still gives you real review and control before anything publishes under your name, since the quality bar doesn't relax just because a machine is doing the writing.
## Related
- [What HarperFlow does](https://www.harperflow.io/docs/what-harperflow-does)
- [Your free trial](https://www.harperflow.io/docs/your-free-trial)
- [How autopilot works](https://www.harperflow.io/docs/how-autopilot-works)
- [Topical authority and content clusters](https://www.harperflow.io/docs/topical-authority-and-clusters)
---
---
Source: https://www.harperflow.io/docs/geo-vs-seo
# GEO vs. SEO: what's actually different
A practical comparison of what search engine optimization and generative engine optimization each reward, and where they overlap.
September 26, 2026
"Is GEO just SEO with a new name?" is one of the first questions anyone asks once they hear the term. The honest answer is no, but they share a goal and a fair amount of groundwork. This chapter lays out exactly where they agree, where they diverge, and why you need both rather than one replacing the other.
## The shared goal
Both disciplines exist to answer the same underlying question: when someone is looking for what you offer, will they find you? SEO answers that by getting a page ranked in a results list. GEO answers it by getting a business named or linked inside an AI-generated answer. The audience's need hasn't changed. The mechanism for satisfying it has split in two.
## Why the split is happening now
The split is happening because the channel itself is shrinking relative to AI answers, not because SEO stopped working. Gartner projects that traditional search engine volume will drop by roughly 25% by 2026 as users shift toward AI chatbots and other virtual agents, a forecast the firm frames explicitly as a prediction rather than a measured fact. Separately, studies of Google search behavior from SparkToro, reported by Search Engine Land, put the current share of Google searches that end without any click to a website at roughly 60 to 68%. Fewer of the searches that do happen result in a visit at all.
Buyer behavior backs this up qualitatively too. In G2's research on B2B software buyers, roughly one in three reported buying from a vendor they'd never previously heard of, surfaced to them by an AI tool rather than by a familiar search result. That's a discovery pattern SEO alone was never built to produce.
## What actually differs
| | Classic SEO | GEO |
| --- | --- | --- |
| What you're optimizing for | Rank position in a results list | Being named or linked inside a generated answer |
| Core mechanism | Crawling and keyword and link-based ranking | Retrieval-augmented generation: retrieve, then generate a cited answer |
| Rewards | Backlink volume, keyword density, domain authority | Semantic clarity, self-contained passages, third-party consensus |
| Where credit shows up | Your own page, in the results list | Often a citation on someone else's page, feeding an AI's answer |
| Success metric | Rankings, organic clicks | Citations, AI-referred visits, brand mentions across the web |
To make the contrast concrete: a classic SEO comparison page might repeat a target keyword phrase in every subheading and chase backlinks through "10 best X tools" listicles. A GEO-oriented version of the same page states each comparison point as a plain, specific sentence, ideally in a table, with a real source behind any number, so any single row can be lifted out and quoted on its own.
## What carries over unchanged
Not everything about SEO becomes irrelevant. A technically healthy, fast, well-organized website still matters: generative engines still need to be able to crawl and parse your site in the first place. Topical relevance still matters. Being a genuine, recognized source in your field still matters, arguably more than before, since both disciplines ultimately reward being a credible answer to a real question. Think of GEO as adding a citation-focused layer on top of a still-necessary SEO foundation, not tearing that foundation out.
> **Note:** If your site currently has no SEO fundamentals in place at all, no clean site structure, no indexable content, no basic technical health, that is the place to start. GEO techniques applied to a page a crawler can't even reach won't help.
## What doesn't carry over
The tactics that carried the most weight in classic SEO, aggressive backlink building, keyword-stuffed headers, thin pages built purely to rank for a term, don't translate to generative engines at all, and some actively work against you. A retrieval system built to ground answers in verifiable, self-contained evidence has no use for a page engineered to rank rather than to inform. The shift in what's rewarded is real, not cosmetic.
## A common resourcing mistake
Once a team learns about GEO, a common reaction is to quietly deprioritize SEO fundamentals altogether, treating the two as a straight either-or decision. That backfires. Technical health and indexability are the precondition for anything to be retrievable in the first place; a generative engine still has to be able to crawl and parse your site before any of its retrieval logic even applies. Dropping SEO basics to chase GEO tactics tends to undermine both at once.
## Why you still need both
Practically, most businesses will keep running SEO fundamentals, site health, indexability, topical structure, while layering GEO-specific practices on top: verifiable claims, answer-first structure, schema, and a genuine trail of third-party mentions. Treating them as a single combined discipline, rather than choosing one over the other, reflects how the two actually interact today.
> **Tip:** A simple test for any piece of content going forward: would this page work as a source citation if an AI pulled a single paragraph out of it, verbatim, with no other context? If the answer is no, it's optimized for the old channel, not the new one.
## Related
- [What is GEO?](https://www.harperflow.io/docs/what-is-geo)
- [How AI search cites content](https://www.harperflow.io/docs/how-ai-search-cites-content)
- [Measuring GEO success](https://www.harperflow.io/docs/measuring-geo-success)
---
---
Source: https://www.harperflow.io/docs/go-live-checklist
# Go-live checklist
What needs to be true before autopilot can start, and what the final setup step does.
September 26, 2026
The last step of setup is "Finish setup & start." Here's exactly what needs to be in place before you reach it, and what actually happens when you click it.
## What you need first
- **A connected, healthy site.** See the [connect guide](https://www.harperflow.io/docs/connect-webflow) for your platform. If your connection has broken since you first set it up, see [Reconnecting and token health](https://www.harperflow.io/docs/reconnecting-and-token-health) before continuing.
- **A reviewed business profile:** the positioning, audience, voice, and topics HarperFlow learned about you in the Learn step, edited if anything was off.
- **A confirmed destination:** a collection or blog chosen and validated, so HarperFlow knows exactly where drafts should land. See [Custom field mapping](https://www.harperflow.io/docs/custom-field-mapping).
- **A plan,** either a free trial or a paid subscription. See [Your free trial](https://www.harperflow.io/docs/your-free-trial).
Authors and promotional callouts are optional and can be skipped or turned off. You can always come back and add them later.
## What "Finish setup & start" actually does
Clicking it runs a short chain. HarperFlow creates your project, picks (or falls back to) keyword and competitor seeds from your profile for discovery, drafts your initial topic clusters, and saves your onboarding profile and collection mapping together. Once this finishes, your workspace moves from "being set up" to "running," ready for you to write, though autopilot itself stays off until you switch it on.
## What happens right after
You're immediately asked whether to turn on autopilot, write a single test article first, or just look around; see [Run your blog on autopilot](https://www.harperflow.io/docs/run-your-blog-on-autopilot). Whichever you pick, nothing ever publishes silently: every article still gets a scheduled publish slot, and you're emailed a summary before anything goes out, a window to review and pull anything you don't want before it becomes live on your site.
> **Note:** If "Finish setup & start" fails partway through, a network hiccup, for example, it's safe to click it again. HarperFlow remembers the project it already created during that attempt and won't create a duplicate; it just resumes from where the chain stopped.
## Related
- [The onboarding wizard, step by step](https://www.harperflow.io/docs/onboarding-wizard)
- [Run your blog on autopilot](https://www.harperflow.io/docs/run-your-blog-on-autopilot)
- [Your free trial](https://www.harperflow.io/docs/your-free-trial)
- [A tour of your dashboard](https://www.harperflow.io/docs/dashboard-tour)
---
---
Source: https://www.harperflow.io/docs/how-ai-search-cites-content
# How AI search cites content
How ChatGPT, Perplexity, and other AI engines actually decide what to retrieve, trust, and cite in an answer.
September 26, 2026
Getting cited by an AI engine isn't a black box you can only guess at. There's a specific mechanism behind it, and understanding that mechanism tells you exactly what to do differently in how you write. This chapter explains how generative engines decide what to pull into an answer and, just as importantly, how that process differs from one engine to the next.
## The core mechanism: retrieval-augmented generation
Most AI answer engines are built on a technique called **retrieval-augmented generation**, or RAG. Instead of answering purely from what a model memorized during training, the system first retrieves a set of relevant documents or passages from an index (or a live web search), then generates its answer grounded in what it retrieved, typically attaching citations back to those sources.
That two-step process matters because it means getting cited is really two separate problems: first, does your content get retrieved at all, and second, once retrieved, does the model choose to quote or link it rather than a competing source. Both steps reward different things than classic search ranking does.
## What the retrieval step rewards
Retrieval systems aren't scanning for keyword density or backlink counts. Research on how these systems select sources points to a few consistent properties:
- **Semantic clarity**, content that plainly states what it means, without requiring the reader (or the retrieval model) to infer the point from context.
- **Structural retrievability**, information organized so a clean, self-contained passage can be lifted out and used on its own, rather than a fact being scattered across several paragraphs.
- **Third-party validation and consensus**, independent agreement about a fact or a business across multiple sources, not just a single first-party claim.
That last point is easy to underestimate. According to analysis from Yext and Profound on how ChatGPT, Perplexity, Gemini, and Claude decide what to cite, roughly 85% of brand mentions in AI answers come from third-party pages, not the brand's own website. In other words, what other credible sites say about you carries more weight in an AI's answer than what you say about yourself.
## Engines don't all behave the same way
One of the most useful, and most overlooked, facts about GEO is that different engines have meaningfully different citation behavior. That same Yext and Profound research found that ChatGPT leans heavily on its training data and high-authority reference sources, with Wikipedia alone accounting for roughly 48% of its top citations, and only searches the live web when it judges that necessary. Perplexity, by contrast, searches the live web on essentially every query and tends to cite around 16 sources per answer.
The overlap between what the two engines actually cite is small: only about 11% of cited domains show up in both ChatGPT's and Perplexity's citations, according to the same analysis. That's a meaningful number for anyone tempted to optimize for a single engine.
> **Note:** The practical implication of that 11% overlap figure is that there is no single trick that gets you cited everywhere. Winning one engine's citation pattern doesn't transfer automatically to the next one.
## Why "be everywhere, credibly" beats "game one engine"
Because engines pull from different mixes of sources, and weigh first-party versus third-party content differently, the only strategy that reliably works across all of them is being consistently and credibly present: accurate information on your own site, plus a genuine trail of third-party mentions, reviews, and references that corroborate it. Trying to reverse-engineer one engine's current citation quirks is a short-lived tactic; broad, real credibility is not.
## Where verifiability fits in
The foundational research behind the term GEO, a study led by researchers at Princeton and published at ACM KDD 2024, tested a range of content-optimization techniques directly against live generative engines and measured the effect on visibility. Its headline finding was that the right optimizations lifted a page's visibility in AI-generated answers by up to 40%, and the single highest-impact techniques it identified were adding relevant statistics, citing credible sources, and including direct quotations, in short, making content verifiable rather than merely well-written.
That finding lines up with the retrieval behavior described above: a system built to ground its answers in retrieved evidence naturally prefers passages that already look like evidence.
> **Tip:** When you're evaluating whether a page is likely to get cited, ask a blunt question: if a fact-checker read only this paragraph, could they verify the claim without leaving the page? If not, that's usually the gap to close first.
## How HarperFlow applies this
HarperFlow writes every article with this retrieval behavior in mind: statements are backed by named sources where a claim needs one, passages are structured so a single section can stand on its own, and the tone stays factual rather than promotional, because that is exactly what a retrieval-and-citation system is built to prefer.
## Related
- [What is GEO?](https://www.harperflow.io/docs/what-is-geo)
- [Structured content principles](https://www.harperflow.io/docs/structured-content-principles)
- [Images and internal linking](https://www.harperflow.io/docs/images-and-internal-linking)
---
---
Source: https://www.harperflow.io/docs/how-autopilot-works
# How autopilot works
How autopilot discovers, drafts, and schedules new articles on your plan's cadence, with a review window before anything publishes.
September 26, 2026
Autopilot is the core of HarperFlow: once it's on, your site gets a steady stream of new articles without you starting each one by hand. This article walks through the loop it runs and what has to be true before it actually fires.
## The loop autopilot runs
Each time autopilot fires, HarperFlow works through the same four steps:
1. **Discover** a keyword worth writing about, based on your niche and what you've already covered.
2. **Draft** a full article: title, body, images, citations, FAQs, and metadata.
3. **Schedule** it with a real future publish time, never "now."
4. **Publish** it automatically once that time arrives, unless you skipped it first.
Between steps 3 and 4 sits the veto window, the only review step in the loop. Nothing needs your sign-off to move forward; something only needs your attention if you want to stop it.
## What has to be true before it runs
Autopilot only dispatches for a project that is:
- **Connected**, with a healthy credential to your site.
- **Mapped**, with a validated field mapping so HarperFlow knows where each part of an article belongs.
- **Taxonomized**, with at least 12 reviewed, active topic clusters.
- **Turned on**, with the autopilot toggle switched on from your Home page. It's off by default after setup; nothing turns it on for you automatically.
Missing any of these doesn't produce an error message anywhere obvious; autopilot just quietly has nothing to do until the gap is closed.
## How often it checks in
Turning autopilot on anchors a weekly cycle to that exact moment: every 7 days from then on, HarperFlow works out how many articles your plan has left and spreads that many writing sessions across the coming week's business days, one article per session. See [Run your blog on autopilot](https://www.harperflow.io/docs/run-your-blog-on-autopilot) for the math behind the count and what turning it off does to already-scheduled drafts.
A session that doesn't fire, because something else is already mid-run for your project, or because of a transient error, isn't lost; it's simply retried the next time HarperFlow checks in rather than skipped for the week.
> **Note:** A project stuck below 12 reviewed topic clusters, or with an unvalidated field mapping, will not error loudly anywhere. Check your project's readiness status if autopilot seems to have gone quiet.
## Related
- [Run your blog on autopilot](https://www.harperflow.io/docs/run-your-blog-on-autopilot)
- [Topic clusters and content types](https://www.harperflow.io/docs/topic-clusters-and-content-types)
- [The veto digest](https://www.harperflow.io/docs/the-veto-digest)
- [Scheduling and spread](https://www.harperflow.io/docs/scheduling-and-spread)
- [Go-live checklist](https://www.harperflow.io/docs/go-live-checklist)
---
---
Source: https://www.harperflow.io/docs/images-and-internal-linking
# Images and internal linking
The featured images, in-body images, and internal links HarperFlow adds to every article automatically.
September 26, 2026
Two things happen to every article without any manual work on your part: it gets imagery, and it gets linked into the rest of your site.
## Images
Every article gets an automatically generated featured image sized for your CMS. Where it fits the content, articles also get a distinct image placed in the body itself, separate from the featured image, so a long article isn't just a wall of text under one header photo. There's no stock-photo search or upload step to do yourself; both are produced, with descriptive alt text included, as part of writing the article.
## Internal linking
When HarperFlow publishes a new article, it automatically looks for opportunities to link it to articles you've already published on the same site, where the topics are genuinely related. Anchors are matched against what the target article is actually about, its title and topic cluster, rather than a keyword match that happens to line up on the surface. If the writing step didn't surface a specific opportunity for a given article, HarperFlow still falls back to matching against an existing article's own keyword or title, so linking doesn't depend entirely on that earlier step succeeding. Links only ever point to your own already-live content; a new article never links to a competitor, and never links to itself.
## Controlling how much
Each project has an internal-link density setting: **Off**, **Light** (up to 2 links), **Standard** (up to 4, the default), or **Rich** (up to 6). Change it from your project settings whenever you want more or fewer links woven per article, and the new setting applies from the next article onward.
> **Note:** Linking runs forward, not backward. A new article can link to your older ones; HarperFlow doesn't go back and edit already-published articles to add links to the new one.
## Related
- [Org Brain and shared memory](https://www.harperflow.io/docs/org-brain-and-shared-memory)
- [Choosing a layout](https://www.harperflow.io/docs/choosing-a-layout)
- [Quality and trust panel](https://www.harperflow.io/docs/quality-and-trust-panel)
---
---
Source: https://www.harperflow.io/docs/link-verification
# Link verification
The two checks HarperFlow runs on an article's sources and links before it ever reaches your schedule.
September 26, 2026
Citing real, working sources is part of what makes HarperFlow's articles trustworthy to both readers and AI search systems. Two separate checks make sure of it.
## Checking sources are real
Before an article is scheduled, HarperFlow checks every source link it plans to cite to confirm it resolves to a real, reachable page, not a dead link or a search-results page that got swept up during research by mistake. A cap limits how many citation links a single article carries, and when good sources are thin, HarperFlow adds real links back in rather than shipping an under-sourced draft.
## A safety gate at publish time
Separately, every article passes through a gate that checks the links themselves for problems your CMS shouldn't carry: leaked search-engine result links, a link whose visible text names your brand but points somewhere you don't own, or a link to a competitor's domain you've told HarperFlow about. Any of these gets automatically corrected or removed. A hallucinated wrong-domain version of your own site, your `.com` cited instead of your real `.io`, for instance, gets quietly rewritten to the domain you actually own, rather than stripped outright.
## When this holds an article back
If that gate catches a real problem, or an article ends up with too few genuine cited sources, HarperFlow holds it in pending review instead of putting it on the schedule automatically. You'll see it as "not on the schedule yet" rather than with a publish time, and it needs a slot from you before it goes anywhere; see [Article states](https://www.harperflow.io/docs/article-states).
> **Tip:** If an article lands in pending review instead of scheduled, check its Quality & trust tab first. Thin sourcing is the most common reason, and the citation count there tells you immediately.
## Related
- [Quality and trust panel](https://www.harperflow.io/docs/quality-and-trust-panel)
- [Article states](https://www.harperflow.io/docs/article-states)
- [Reviewing before publish](https://www.harperflow.io/docs/reviewing-before-publish)
---
---
Source: https://www.harperflow.io/docs/managing-multiple-sites
# Managing multiple sites
How to add, switch between, and remove sites, and what a loyalty discount gets you.
September 26, 2026
You can connect as many sites as you like, even more than one on the same platform.
## Adding another site
Open the site switcher and choose **Add another site**. This opens a brand-new workspace and drops you straight into setup for it, connecting whichever platform and site you choose next. Each site gets its own plan, unless your account is on Agency (see [Accounts, organizations, and sites](https://www.harperflow.io/docs/accounts-orgs-and-sites)).
## Switching between sites
The site switcher lists every workspace you belong to. Picking one swaps your active workspace and reloads the dashboard into that site's own articles, Brain, and settings.
## The loyalty discount
If you already own or admin a workspace with an active paid plan, your next new site automatically gets a **20% discount**, applied for as long as you keep it. There's no code to enter; it's applied at checkout for that site.
## Removing a site
An empty workspace, one you started but never finished connecting a site to, can be removed directly from the site switcher, since it has nothing in it to lose. A site you've actually connected is removed from that site's own settings rather than the switcher.
Deleting your **last** site on a Solo account schedules its plan to cancel at the end of your current billing period. You keep access until then and are never charged again after that. If you still have other active sites, or your account is on Agency, deleting one site doesn't touch billing at all; the rest keep running as normal.
If your plan is billed through Shopify, deleting the site doesn't end that charge. Uninstall the HarperFlow app from your Shopify admin instead; Shopify ends the app's billing when it's uninstalled.
> **Note:** Deleting a site never triggers an immediate charge or an immediate cutoff. At most, it schedules the plan to cancel at the end of the period you've already paid for, and only when it was genuinely your last active site.
## Related
- [Accounts, organizations, and sites](https://www.harperflow.io/docs/accounts-orgs-and-sites)
- [Your free trial](https://www.harperflow.io/docs/your-free-trial)
---
---
Source: https://www.harperflow.io/docs/measuring-geo-success
# Measuring GEO success
What to actually track for generative engine optimization, since AI citations don't show up in a standard analytics dashboard.
September 26, 2026
Traditional SEO has a well-worn measurement playbook: rankings, organic clicks, backlinks. GEO doesn't have an equally mature one yet, largely because the engines themselves don't expose a citations dashboard for you to check. That makes measurement the part of GEO most teams skip entirely. This chapter covers what's actually available to track, and why it's worth the extra effort even without a single clean number to watch.
## Start by simply asking the engines
The single most direct measurement available to you doesn't require any tooling. Ask ChatGPT, Perplexity, and Google's AI answers a real question in your category, the kind an actual prospect would ask, and see what comes back. Are you mentioned at all? Is what's said about you accurate? Who is mentioned instead of you? Repeating this periodically, for the handful of questions that matter most to your business, is a legitimate and underused audit.
This matters because absence here doesn't show up as a low ranking the way it would in classic search. It shows up as nothing: you simply aren't part of the answer. A recurring manual audit is currently the most reliable way to see that gap directly.
## Track AI-referred traffic as its own category
Where your analytics can help is on the other side of a citation: once someone clicks through from an AI tool's answer to your site, that visit is measurable like any other, provided you can distinguish it from generic organic traffic. Most modern analytics setups can isolate referral traffic from AI tools as its own segment. Once you can see that segment, its behavior is worth watching closely, because there's real evidence it converts differently.
In a 12-month analysis across 94 e-commerce brands, referral traffic from AI tools like ChatGPT converted roughly 31% higher than non-branded organic search traffic, according to research published by ALM Corp. The likely explanation is straightforward: a visitor who arrived because an AI already summarized and effectively endorsed you is further along in their decision than someone who clicked an unfamiliar link off a results page.
## Why consistency is a metric worth watching, not just an input
Publishing cadence is usually treated as an input to a content program, not something you measure as an outcome. It's worth tracking directly. Businesses that maintain a consistent blogging cadence generate roughly 67% more leads than those that don't, and consistent publishers see content ROI roughly 13 times higher than sporadic ones, according to content-marketing benchmarks compiled by HubSpot and Ranktracker. Since GEO rewards that same sustained cadence, fresh content and ongoing topical coverage, tracking whether you actually kept up your planned publishing schedule is itself a leading indicator worth reviewing monthly, well before you get to citation-level results.
## The opportunity gap is itself worth tracking
One more figure is useful less as a metric to hit and more as context for how much runway is left: roughly 70% of marketers say answer-engine and generative-engine optimization will be critical to their work within the next one to three years, yet only around 20% report having actually started, according to an AEO statistics roundup published by Omnibound. If that gap is accurate, most of your competitors have not yet built the citation history and third-party consensus that compounds over time. Measuring your own progress against a mostly empty field is a different exercise than measuring it against a saturated one.
> **Note:** None of the figures above are a substitute for checking your own numbers. They describe industry-wide patterns, not a guarantee for any specific business. Use them to calibrate urgency, not to forecast your own results.
## Building a simple measurement habit
In practice, a workable measurement routine for GEO doesn't need to be elaborate:
1. **Monthly:** ask the major AI engines your top three to five category questions and note what changed.
2. **Monthly:** check whether your planned publishing cadence actually happened, and what got skipped.
3. **Ongoing:** segment AI-referral traffic in analytics where your tooling supports it, and watch its conversion rate against your other channels.
4. **Quarterly:** revisit older articles for freshness, since stale content loses citations over time regardless of how it performed originally.
> **Tip:** Keep a simple running log of the manual AI-engine audits: question asked, date, what each engine said. A single snapshot tells you little; a few months of them shows you a trend, which is the actual signal you're after.
## How HarperFlow applies this
Every HarperFlow article carries a quality score and a verification log you can inspect individually, and the Brain view rolls up published article count, total words live, sources cited, and a rolling weekly publishing trend across your site, so the cadence and coverage side of this measurement routine is visible without separate tracking.
## Related
- [Quality and trust panel](https://www.harperflow.io/docs/quality-and-trust-panel)
- [Org Brain and shared memory](https://www.harperflow.io/docs/org-brain-and-shared-memory)
- [GEO vs. SEO: what's actually different](https://www.harperflow.io/docs/geo-vs-seo)
---
---
Source: https://www.harperflow.io/docs/on-demand-runs
# On-demand runs
What happens when you trigger a run yourself instead of waiting for autopilot's next scheduled tick.
September 26, 2026
Autopilot runs on its own cadence, but you can also trigger a run yourself at any moment. HarperFlow calls this an on-demand run, and it's the same writing pipeline autopilot uses, just started by a click instead of a clock.
## Single article or a full batch
You choose the scope when you trigger a run:
- **Write one** produces exactly one article right now, on a topic HarperFlow discovers, or one you supply yourself; see [Topic runs](https://www.harperflow.io/docs/topic-runs).
- **Write everything remaining** uses up whatever is left of your plan's quota for the current billing period in one request, spread across the rest of that period rather than published all at once.
## It still gets a review window
An on-demand article isn't published the instant it's written. It goes through the exact same scheduling step as an autopilot article: a real future publish slot, with the veto window open until that time arrives. Requesting a run yourself doesn't skip the review step; it just decides when the writing starts.
## Your free trial runs on this
If you're on a free trial, on-demand runs are how you experience HarperFlow before committing to autopilot: your trial period and article allowance are delivered through on-demand requests you make yourself, at your own pace.
## Cancelling a run
Right after you request a run, there's a short window, about two minutes, where you can cancel it outright with nothing charged against your quota. After that window closes, the run is already being written, and any articles it produces will count against your budget as they land, though you can still skip any of them individually before they publish.
> **Note:** Even a single-article, on-demand request doesn't publish immediately. It produces a scheduled draft with its own publish time and veto window, exactly like an autopilot article does.
## Related
- [Topic runs](https://www.harperflow.io/docs/topic-runs)
- [Scheduling and spread](https://www.harperflow.io/docs/scheduling-and-spread)
- [Article states](https://www.harperflow.io/docs/article-states)
- [Your free trial](https://www.harperflow.io/docs/your-free-trial)
---
---
Source: https://www.harperflow.io/docs/onboarding-wizard
# The onboarding wizard, step by step
What each of the seven setup steps does, from picking a plan to going live.
September 26, 2026
Each new site goes through setup once. Shopify users begin by installing [HarperFlow from the Shopify App Store](https://apps.shopify.com/harperflow); Shopify handles trial and subscription approval. The overview below describes the setup stages; available options depend on your platform. It's seven steps, and you can leave and come back: nothing publishes until you finish the last one.
## Step 1: Plan and account type
You choose Solo or Agency (see [Accounts, organizations, and sites](https://www.harperflow.io/docs/accounts-orgs-and-sites)) and see the shape of the deal: one connected site, a monthly article volume, and a schedule that spaces articles out automatically. If more than one platform is available to you, you'll also pick which one you're publishing to here.
## Step 2: Connect
Link the site itself. Each platform connects a little differently: see [Connect your Webflow site](https://www.harperflow.io/docs/connect-webflow), [Connect your WordPress site](https://www.harperflow.io/docs/connect-wordpress), [Connect your Shopify store](https://www.harperflow.io/docs/connect-shopify). Wix support is [in development](https://www.harperflow.io/docs/connect-wix).
## Step 3: Learn your site
HarperFlow reads your connected site and drafts a business profile for you to review, covering positioning, audience, voice, topics, and the article formats it should write. Nothing here is final. Edit any of it, or add and remove topics and formats, before continuing. If the read fails or your site has no content yet, you can fill the profile in yourself instead.
## Step 4: Callouts
HarperFlow drafts a few promotional call-to-action blocks based on your profile, things like "explore \[product]" or "talk to the team," that can appear inside articles. Edit the wording, or regenerate them.
## Step 5: Authors
Optionally add up to three bylines, people or companies, that HarperFlow attributes articles to. This is entirely optional. Turn it off and articles publish without an author section.
## Step 6: Confirm your destination
On Webflow, choose a new collection, extend an existing collection, or map its fields. See [Custom field mapping](https://www.harperflow.io/docs/custom-field-mapping) for those Webflow options. On Shopify, choose the blog that should receive your articles; your store theme controls the article layout. Follow the destination options shown for your connected platform.
## Step 7: Go live
Choose your article volume and review the trial or subscription terms. Shopify users approve recurring charges in Shopify; Webflow and WordPress use the web subscription flow. HarperFlow then finishes setup. It creates your project, drafts your topic clusters, and wires up the collection you chose. Autopilot itself stays off until you turn it on, either from the prompt shown right after setup or later from your Home page. See [Your free trial](https://www.harperflow.io/docs/your-free-trial), [Run your blog on autopilot](https://www.harperflow.io/docs/run-your-blog-on-autopilot), and the [go-live checklist](https://www.harperflow.io/docs/go-live-checklist) for what happens next.
> **Note:** Your answers are saved to this browser as you go, not to your account. If you use "Finish later" and come back on the same device and browser, you'll pick up right where you left off, but switching devices or clearing your browser storage partway through means redoing the earlier steps.
## Related
- [Your free trial](https://www.harperflow.io/docs/your-free-trial)
- [Run your blog on autopilot](https://www.harperflow.io/docs/run-your-blog-on-autopilot)
- [Go-live checklist](https://www.harperflow.io/docs/go-live-checklist)
- [Custom field mapping](https://www.harperflow.io/docs/custom-field-mapping)
---
---
Source: https://www.harperflow.io/docs/org-brain-and-shared-memory
# Org Brain and shared memory
The read-only Brain page showing your brand voice, topic coverage, and content growth, plus what shared memory does across sites.
September 26, 2026
The Brain page is where HarperFlow shows you what it knows about your site and how your content library is growing. For accounts with more than one site, it's also where shared memory lives.
## What the page shows
Your Brain page pulls together:
- **Brand voice**: the profile described in [Editing your brand voice](https://www.harperflow.io/docs/editing-your-brand-voice).
- **Topic coverage**: your reviewed clusters, how many articles exist in each, and which ones are still gaps.
- **Internal links**: how many links have been woven into your published articles, and your current density setting.
- **Growth**: published article count, total words live, sources cited, and a rolling weekly publishing trend.
## Writing directly from a gap
A cluster with zero articles shows up as a gap. You can seed a new run against it directly from the coverage map instead of waiting for autopilot or discovery to pick it up on its own.
## Shared memory across sites
If your organization manages more than one site, a **shared memory** toggle controls how much they act as one connected program instead of separate ones. Turned on:
- Every site shares a single, org-level brand voice instead of each keeping its own.
- Topic memory spans all the org's sites, so one site never rewrites a topic another site already covered.
- Recent skips and take-downs on any site inform what gets written on every other site in the org.
## Agencies: a portfolio view
An Agency account managing several organizations sees a further rollup one level up: total coverage and growth across every organization it manages, plus a breakdown per site, gated to people with account-level access.
> **Warning:** Turning shared memory on or off changes behavior for every site in the organization at once, not just the one you're currently looking at.
## Related
- [Editing your brand voice](https://www.harperflow.io/docs/editing-your-brand-voice)
- [Topic clusters and content types](https://www.harperflow.io/docs/topic-clusters-and-content-types)
- [Accounts, organizations, and sites](https://www.harperflow.io/docs/accounts-orgs-and-sites)
- [Managing multiple sites](https://www.harperflow.io/docs/managing-multiple-sites)
---
---
Source: https://www.harperflow.io/docs/plans-and-pricing
# Plans and pricing
Choose your article volume, with separate Webflow/WordPress and Shopify pricing.
September 26, 2026
HarperFlow includes the full publishing pipeline at every article volume. Choose **5–30 articles per month** for a single site. Your publishing platform determines the pricing and billing provider.
## Webflow and WordPress
Webflow and WordPress subscriptions are billed through Stripe. The USD price per article is $12.90 minus $0.10 for each article in your monthly package.
| Articles per month | Monthly price (USD) |
| --- | --- |
| 5 | $62 |
| 10 | $119 |
| 20 | $218 |
| 30 | $297 |
Every whole-number volume between 5 and 30 is available. The [pricing page](https://www.harperflow.io/pricing) shows the web subscription options. The web trial lasts 7 days and includes 3 articles; payment details are required at checkout.
## Shopify
Shopify uses its own pricing curve and **Shopify Billing**. The web pricing formula above does not apply.
| Articles per month | Monthly price (USD) |
| --- | --- |
| 5 | $39.90 |
| 10 | $74.90 |
| 20 | $134.90 |
| 30 | $184.90 |
The [Shopify App Store listing](https://apps.shopify.com/harperflow) offers a **7-day free trial**. Review the exact trial terms and recurring charge in Shopify before approving your subscription. Start by [installing and connecting your store](https://www.harperflow.io/docs/connect-shopify).
## Included at every volume
Research, source citations, brand voice, structured articles, images, scheduling, and direct publishing are included. You can veto scheduled articles before they go live. Platform-specific implementation differs: Shopify uses your blog and theme; Webflow supports mapped CMS fields.
## Change or cancel
Use HarperFlow's **Plan** page to change article volume. Shopify merchants approve changes through Shopify; Webflow and WordPress customers follow the web billing flow. See [changing your volume](https://www.harperflow.io/docs/upgrading-and-downgrading) and [canceling](https://www.harperflow.io/docs/canceling-your-subscription).
## Agencies and multiple sites
Agency and Enterprise arrangements start at $1,020 per month for 120 articles. [Contact the founder](https://www.harperflow.io/contact) for a multi-site arrangement. Do not assume a Shopify App Store subscription uses consolidated Stripe agency billing.
---
---
Source: https://www.harperflow.io/docs/publish-failures-error-reference
# Publish failure error reference
Every error message HarperFlow shows you, grouped by area, with what to do about each one.
September 26, 2026
Every error HarperFlow shows you is translated into plain language before it reaches your screen. You should never see a raw technical error. This page lists those messages, grouped by area, with what each one means and what to do.
## Connection and site health
| What you'll see | What to do |
| --- | --- |
| Your connection looks expired or invalid | Reconnect your site from the Your Site page. |
| HarperFlow has lost the connection to your site | Reconnect your site from the Your Site page. |
| Your token cannot see any sites | It's likely scoped to a whole workspace instead of one site. Generate a site-scoped token instead. |
| HarperFlow couldn't reach your site just now | Retry. Nothing was saved or lost. |
| HarperFlow couldn't find readable content on your site yet | Add some content to your site, or fill in the basics yourself. |
| HarperFlow couldn't read your site automatically | Retry, or fill in the basics yourself. Either works. |
## Collection and field mapping
| What you'll see | What to do |
| --- | --- |
| HarperFlow hit a snag reading your collection's fields | Retry. If it keeps failing, pick a different collection or map fields manually. |
| Your collection has required fields HarperFlow can't fill | Make those fields optional in your CMS, or let HarperFlow create a fresh collection. |
| A collection with that name already exists | Recently deleted names can stay reserved briefly. Try a different name. |
| That collection was deleted from your site | Refresh the page and pick one that still exists. |
| HarperFlow can't match your collection to its saved mapping | Usually means the collection changed. Re-run the collection step in setup. |
| HarperFlow couldn't sync your collection's topic fields | Retry. If it keeps failing, let HarperFlow create a fresh collection. |
| Your collection mapping isn't finished | Complete the collection step in setup. |
| Your mapping needs a quick validation | Re-run provisioning from the collection step in setup. |
## Billing and plan
| What you'll see | What to do |
| --- | --- |
| This workspace already has an active plan | Use Manage plan to switch tiers instead of subscribing again. |
| You've used all your articles for this period | Upgrade your plan, or wait for next month's reset. |
| This needs an active trial | Check your Plan page for your current trial status. |
## Runs, autopilot, and scheduling
| What you'll see | What to do |
| --- | --- |
| This run already finished | Nothing to do. It completed. |
| A run is already in progress | Check the run banner for its current status. |
| HarperFlow couldn't create your run | Retry. Your quota wasn't touched. |
| HarperFlow couldn't start the writing pipeline | Retry in a moment. Nothing was charged. |
| This project isn't active | Check your project settings. |
| Writing is temporarily paused | Try again shortly. |
| This workspace is on a break | Resume it from the Take a Break card to generate articles. |
| Too many runs in flight right now | Try again in a moment. |
| Autopilot isn't turned on for this project | Turn it on from the project's autopilot settings. |
## Topics and setup
| What you'll see | What to do |
| --- | --- |
| You need at least 12 reviewed topic clusters before writing | Finish the topic-cluster step in onboarding. |
| HarperFlow couldn't load your topic clusters | Retry. |
| HarperFlow couldn't load your workspace | Retry. |
| HarperFlow couldn't save your setup | Retry. Nothing you'd already entered is lost. |
> **Note:** Anything not on this list falls back to a generic message: "We couldn't load this right now. Retry — your data is safe." That's a deliberate default for the unexpected, not a sign something is broken beyond a retry.
> **Tip:** If a retry doesn't resolve it and the issue isn't listed here, use [Contacting support](https://www.harperflow.io/docs/contacting-support). Include what you were doing when it happened.
## Related
- [Reconnecting and token health](https://www.harperflow.io/docs/reconnecting-and-token-health)
- [Sitemap diagnosis](https://www.harperflow.io/docs/sitemap-diagnosis)
- [Custom field mapping](https://www.harperflow.io/docs/custom-field-mapping)
- [Contacting support](https://www.harperflow.io/docs/contacting-support)
---
---
Source: https://www.harperflow.io/docs/publish-in-a-second-language
# Publish in a second language
Add a secondary publishing language, pick its search market, and control what share of your autopilot articles are written in it.
September 26, 2026
HarperFlow can write part of your autopilot articles in a second language, researched against that language's own search market. You control all of it from one card on **Your site**.
## Turning it on
On **Your site**, find the **Second publishing language** card and tick **Enable**. Three choices appear:
- **Language** — any supported publishing language other than your site's primary one.
- **Search market** — the country whose search demand topic discovery should use for this language. Leave it on the language default if you're not sure; pick a specific market when your audience is concentrated there (for example French with the France market, or Spanish with Mexico rather than Spain).
- **Share of articles** — a slider from 10% to 50%. This is the portion of your autopilot articles written in the secondary language.
Save, and the next autopilot run takes it into account. No restart or reconnection is needed.
## How the share works
The share is honored **over time, not per week**. Before each run, HarperFlow looks at what it actually published for your site in the trailing thirty days and writes the next article in whichever language is furthest behind its target. At 25%, roughly one article in four ends up in your secondary language; a given week can naturally be all-primary or all-secondary.
Two things never switch language:
- **Topics you type yourself** are written in your primary language by default — but once a second language is enabled, the "Have a topic in mind?" form on Home grows a language picker, so you can write any manual article in either language.
- **Existing scheduled articles** keep the language they were written in.
## What each language gets
A secondary-language article is not a translation. It gets the full pipeline in that language: topic discovery against the market you chose, sources evaluated for that language, writing, formatting, and publishing. Where few native-language sources exist for a topic, HarperFlow may lean on carefully verified translated evidence rather than refusing to write — the sourcing bar stays the same.
## Webflow sites: locale check at save time
On Webflow, articles in a secondary language publish into the matching **Webflow Localization locale**, under its own URL prefix (for example `/fr/...`). Because of that, HarperFlow checks your site's locale setup when you save: if the language you picked isn't enabled in your Webflow Localization settings, the save is rejected with the list of locales your site actually has. Enable the locale in Webflow first (Site settings → Localization), then save again.
WordPress and Shopify sites don't need any locale setup — articles publish normally and carry the right language metadata.
## Turning it off
Untick **Enable** and save. Discovery and writing return to your primary language only. Already-written secondary-language articles are unaffected: scheduled ones still publish, live ones stay live.
---
---
Source: https://www.harperflow.io/docs/quality-and-trust-panel
# Quality and trust panel
The per-article scorecard showing how HarperFlow checked its own writing before you ever see it.
September 26, 2026
Every article has a **Quality & trust** tab showing an automatic self-check HarperFlow runs after writing it. It's meant to answer a simple question: how much of this can you trust without reading every word yourself? This article covers what's on it.
## Six dimensions, one verdict
HarperFlow scores each article across six dimensions: factual grounding, citation quality, profile fidelity (how well it matches your brand), GEO optimization, writing quality, and formatting and semantics. Those roll up into an overall score and a verdict of pass, warn, or fail, along with a separate score for the article's generated image.
## Sources and facts
The panel shows how many sources the article cites and how many specific facts are grounded in them, so "well-researched" is something you can check rather than take on faith. You'll find this tab on the article's own detail view, right alongside its body, metadata, and citations.
## Verification summary
Alongside the score, you'll see a short verification summary: how many numeric or factual claims were checked, how many were confirmed as accurate, how many were corrected to match their source, and how many couldn't be verified at all. [Link verification](https://www.harperflow.io/docs/link-verification) covers this in more depth, along with how sources are checked before an article is even scheduled.
## What a low score means for scheduling
A warn or fail verdict doesn't hold an article back on its own. It runs after the article's already scheduled, and shows up as a **Review suggested** flag in your Articles queue so you know which upcoming articles are worth a closer read during your veto window. Whether to skip it is still your call.
> **Note:** The audit runs in the background right after an article is created, so it never delays scheduling. A verdict can take a little while to show up on a brand-new article.
## Related
- [Link verification](https://www.harperflow.io/docs/link-verification)
- [Reviewing before publish](https://www.harperflow.io/docs/reviewing-before-publish)
- [Article states](https://www.harperflow.io/docs/article-states)
---
---
Source: https://www.harperflow.io/docs/reconnecting-and-token-health
# Reconnecting and token health
What happens when your site connection breaks, and how to reconnect per platform.
September 26, 2026
Connections can break. A token gets revoked, a permission gets removed, or a password rotates. Here's how HarperFlow tells you, and what to do about it.
## How you'll know
When HarperFlow can't reach your site, a red banner appears across your dashboard: "We can't reach your \[platform] site, reconnect to resume publishing." Publishing pauses until the connection is fixed. Nothing is lost while it's down.
## Reconnecting on Webflow
Webflow's connection is authorization-based, not a password you re-type. The banner's "Reconnect with Webflow" link takes you through Webflow's login flow again for this specific site, rather than asking you to paste a new token.
## Reconnecting on WordPress
Go to your Site page and reconnect with a fresh Application Password, using the same three fields (site URL, username, Application Password) as the original connect step.
## Reconnecting on Shopify
If you're using HarperFlow from inside your Shopify admin, click **Reconnect** on the banner. It re-establishes the connection automatically without leaving Shopify. If that's not available, reopen HarperFlow from your Shopify admin, or re-approve HarperFlow's permissions there if Shopify prompts you to.
## Wix availability
Wix support is in development and is not currently available for publishing.
## If it stays broken
Contact support if reconnecting does not restore publishing. A publishing pause is not a cancellation request: check your subscription status before assuming billing has stopped. Shopify subscriptions and outstanding charges are managed through Shopify. See [Canceling your subscription](https://www.harperflow.io/docs/canceling-your-subscription) for the steps that apply to your platform.
## Related
- [Connect your Webflow site](https://www.harperflow.io/docs/connect-webflow)
- [Connect your WordPress site](https://www.harperflow.io/docs/connect-wordpress)
- [Connect your Shopify store](https://www.harperflow.io/docs/connect-shopify)
---
---
Source: https://www.harperflow.io/docs/reviewing-before-publish
# Reviewing before publish
How to review scheduled articles before they publish, now that there's no separate approval step to click through.
September 26, 2026
There's no "approve" button in HarperFlow. Every article is already scheduled the moment it's written, and reviewing it means deciding, any time before its publish slot, whether to let it go live or stop it.
## Where to look
Your Articles page is segmented into **Upcoming**, **Published**, and **Skipped**. Upcoming is where the review actually happens: every scheduled article, and any rare pending-review article, with its publish time shown right on the row. Switching between the three tabs tells you, at a glance, how much of your recent output is still awaiting a decision versus already settled one way or the other.
## What you can do
On any upcoming article, these three actions cover everything you can do without leaving the queue:
- **Skip this one**, the red primary action, which stops it from publishing and optionally asks for a line of feedback that tunes future articles.
- **Change the time**, if the slot itself just needs to move.
- **Publish now**, if you'd rather not wait for its slot at all.
## Signals worth a closer look
Two things call out an article that deserves extra attention before its slot arrives: a **Review suggested** tag when its automatic quality check came back below the bar, and an amber "last chance to skip" chip once it's within 24 hours of publishing.
## Changing your mind
Skipped an article and reconsidered? A **Restore** action on any skipped, or taken-down, article brings it back to pending review, ready for a new slot, exactly as if it had just been written.
> **Warning:** Skip only works before an article publishes. Once it's live on your site, use [Take this down](https://www.harperflow.io/docs/unpublish-and-take-down) instead; it's the equivalent action for something that already went out.
## Related
- [The veto digest](https://www.harperflow.io/docs/the-veto-digest)
- [Article states](https://www.harperflow.io/docs/article-states)
- [Unpublish and take-down](https://www.harperflow.io/docs/unpublish-and-take-down)
- [Quality and trust panel](https://www.harperflow.io/docs/quality-and-trust-panel)
---
---
Source: https://www.harperflow.io/docs/run-your-blog-on-autopilot
# Run your blog on autopilot
How the autopilot toggle plans your week, the math behind the count, and what happens when you turn it off.
September 26, 2026
**Run My Blog On Autopilot** is the switch on your Home page that puts publishing on a recurring, hands-off cadence. This article covers what the switch actually does week to week; see [How autopilot works](https://www.harperflow.io/docs/how-autopilot-works) for the discover-draft-schedule-publish loop it runs underneath.
## Turning it on
Nobody's autopilot is on by default. After setup finishes, you get a one-time prompt asking whether to turn it on, write a single test article first, or just look around; if you skip that, the same switch sits on your Home page whenever you're ready. Turning it on requires an active paid plan or trial; a brand-new, unpaid workspace won't show a working switch yet.
## How many articles you get each week
The moment you turn it on, that instant becomes your weekly anchor: every 7 days from then on is one cycle. At the start of each cycle, HarperFlow looks at how many articles your plan has left and how many weeks remain until your quota resets, and divides one by the other, rounded to a whole article, to decide that week's count.
For example, 12 articles left with 4 weeks to go plans about 3 for this week. If you also write 2 articles yourself that week, nothing about the current week's plan changes retroactively, but the following week's math sees fewer articles remaining and automatically plans a smaller batch to match. You never get more than your plan allows; a slow week just means a lighter one, not a skipped one.
Once the week's count is set, HarperFlow spreads that many writing sessions across your business days, one article per session. Each article it writes is then scheduled to publish about a day later, never the moment it's written, so there's always room for the veto window before anything goes live.
## Nothing publishes without a review window
Every article autopilot writes lands as a **scheduled** draft with a real publish time, and you get a veto email before that time arrives. See [The veto digest](https://www.harperflow.io/docs/the-veto-digest) and [Article states](https://www.harperflow.io/docs/article-states).
## Turning it off
Switching the toggle off stops planning new articles and asks what to do with drafts already on the schedule:
- **Keep them scheduled.** They publish as planned unless you skip them individually first.
- **Unschedule them.** They move to Articles → Skipped, where you can restore or publish them anytime.
> **Warning:** Credits already spent writing a scheduled draft aren't refunded either way, even if you choose to unschedule it.
## If you're on a free trial
Trials get their 3 articles planned across the trial week rather than the usual monthly cycle, so you see the full loop, discovery through a scheduled publish, before you ever pay.
## When your quota runs out
If your plan's articles are already used up for the current period, autopilot plans nothing new until your quota resets. The switch stays on and picks back up automatically once it does; nothing needs re-enabling.
## Related
- [How autopilot works](https://www.harperflow.io/docs/how-autopilot-works)
- [The Usage tab](https://www.harperflow.io/docs/the-usage-tab)
- [The veto digest](https://www.harperflow.io/docs/the-veto-digest)
- [Article states](https://www.harperflow.io/docs/article-states)
- [Your free trial](https://www.harperflow.io/docs/your-free-trial)
---
---
Source: https://www.harperflow.io/docs/scheduling-and-spread
# Scheduling and spread
How HarperFlow spreads freshly written articles across future publish times instead of releasing them all at once.
September 26, 2026
No article HarperFlow writes ever publishes the moment it's created. Every one gets a real, specific slot in the future, and this article covers how those slots get chosen.
## How a slot is picked
Publish times land within your plan's cadence: the days of the week and business-hours window your schedule is set to, in your project's own timezone. Within that window, HarperFlow adds a bit of randomness to the exact minute so a batch of articles doesn't all land at the identical time.
A sparser cadence (say, two days a week) spreads across a longer window than a dense one (every weekday), so there are always enough eligible days to spread a batch across without doubling up on any one of them.
## Runs don't collide with each other
If you trigger a run while another batch is already scheduled, the new slots start after your most recently scheduled article for that project, not from the moment the new run began. Back-to-back or overlapping runs cascade forward in time instead of piling multiple articles onto the same slot.
## Autopilot ticks vs. a manual batch
Autopilot plans on its own weekly cycle, anchored to the day you turned it on: each week it writes roughly your plan's remaining articles divided by however many weeks are left before your quota resets, one article at a time, so a single week doesn't spend the whole period's budget at once. See [Run your blog on autopilot](https://www.harperflow.io/docs/run-your-blog-on-autopilot) for the exact math. A manual "write everything remaining" batch works differently: it spreads across whatever is left of your current billing period (at least a week, capped at 45 days) instead of a fixed week, since it represents the rest of what you're already paying for this period.
> **Tip:** If you move one article's publish time by hand, the rest of the batch stays where it was. Reschedule any others individually from the Articles queue if you want to change more than one.
## Related
- [How autopilot works](https://www.harperflow.io/docs/how-autopilot-works)
- [Run your blog on autopilot](https://www.harperflow.io/docs/run-your-blog-on-autopilot)
- [Article states](https://www.harperflow.io/docs/article-states)
- [On-demand runs](https://www.harperflow.io/docs/on-demand-runs)
---
---
Source: https://www.harperflow.io/docs/security-and-account-deletion
# Security and account deletion
Setting up two-factor sign-in, and how to remove a member or close a workspace.
September 26, 2026
Two ways to lock things down: securing how you sign in, and removing access or workspaces you no longer need. Neither one requires digging through settings you'll never touch again.
## Two-factor authentication
From your Profile page, you can add a second sign-in step on top of your password:
- **An authenticator app**: scan a QR code with an app like Google Authenticator or Authy, then confirm with the 6-digit code it shows.
- **A passkey**: use Face ID, Touch ID, or a hardware security key instead of a code.
Once you've enrolled either one, HarperFlow requires that second step at sign-in before you can reach the rest of the app. A password alone is no longer enough. If you ever lose access to your authenticator app or passkey, [contact support](https://www.harperflow.io/docs/contacting-support): the team can clear your enrollment as a lockout backstop so you can sign in and set it up again.
## Removing a teammate
Covered in more detail in [Team members and roles](https://www.harperflow.io/docs/team-members-and-roles), but worth repeating here: removing someone is a real access change, not just a cosmetic one. It takes effect immediately, cutting their access across every workspace under the account and signing them out, rather than just hiding their row on the Members page. You can always invite them back later if the removal was a mistake or their role changes.
## Closing a workspace
A workspace (organization) with no connected site can be removed entirely. If it still has a site attached, disconnect or delete that site first. You also can't remove the workspace you currently have open. Switch to another one first, then remove it from there. Removing a workspace cancels any subscription still attached to it and clears its data.
> **Warning:** Closing a workspace is permanent once submitted. Make sure you've disconnected its site and are certain before you remove it. There's no undo.
## Related
- [Team members and roles](https://www.harperflow.io/docs/team-members-and-roles)
- [Managing multiple sites](https://www.harperflow.io/docs/managing-multiple-sites)
---
---
Source: https://www.harperflow.io/docs/seo-and-schema-fields
# SEO and schema fields
The structured-data, social-preview and citation fields HarperFlow writes into your collection, and how to bind them in Webflow.
September 26, 2026
Alongside the article itself, HarperFlow computes a set of fields that exist purely to make each post easier for search engines and AI answer engines to read. They're written into your collection like any other field, so **you can bind them in the Designer yourself** — no app, no custom code permissions, nothing to install.
## What gets written
| Field | What it holds |
| --- | --- |
| **Article Schema** | The article's `Article` structured data, as JSON. |
| **OG Title** | The title social platforms and AI crawlers should show. |
| **OG Description** | The one-line description they should show with it. |
| **OG Image** | The image they should show with it. |
| **Canonical URL** | The article's own live URL, so duplicates elsewhere don't compete with it. |
| **Word Count** | How many words the published article actually runs to. |
| **Sources** | The outbound citations the article links to, as a list. |
| **Related Articles** | The links to your own posts that the article weaves in. |
| **Author Snapshot** | The byline, as plain text. |
| **Author Name** | The byline name on its own, without the role. The article's structured data uses it as the author. |
| **TLDR** | Your key takeaways as a single paragraph, for a lede under the headline. |
Every one is optional. If a field isn't in your collection, or an article genuinely has nothing to put in it — no featured image, no outbound citations — HarperFlow leaves it out rather than writing something empty. HarperFlow adds **Author Name** only when your collection has room for it under Webflow's 60-field limit.
## Binding Article Schema in Webflow
Article Schema holds **bare JSON**, deliberately without a `
```
4. Put your cursor between the tags and add the **Article Schema** field from the CMS field picker.
5. Publish.
Every article then serves its own valid structured data in the page's raw HTML — which matters, because most AI crawlers read raw HTML and never run JavaScript.
## Binding the rest
- **OG Title / OG Description / OG Image** — in the collection template's page settings, under Open Graph, bind each to its matching field.
- **Canonical URL** — add it to a `` tag in the page's custom code, or bind it in page settings if your plan includes that.
- **Sources**, **Related Articles**, **TLDR**, **Author Snapshot**, **Author Name**, **Word Count** — ordinary fields; drop them anywhere in your template and style them however you like.
## If you use the HarperFlow Connector
The Connector applies a template that binds all of this for you in a click, so you don't need to do any of the above. These fields exist so that everything still works if you'd rather set your template up by hand.
## Related
- [Custom field mapping](https://www.harperflow.io/docs/custom-field-mapping)
- [Article URL and Post Markdown fields](https://www.harperflow.io/docs/article-url-and-post-markdown-fields)
- [Choosing a layout](https://www.harperflow.io/docs/choosing-a-layout)
---
---
Source: https://www.harperflow.io/docs/shopify-and-wix-connection-issues
# Shopify connection issues and Wix availability
Reconnect Shopify through the App Store flow; Wix remains in development.
September 26, 2026
## Shopify connection issues
The standard Shopify connection comes from installing the HarperFlow app. Uninstalling it revokes access to the store, so HarperFlow cannot publish new articles through that connection.
To reconnect, install or reopen [HarperFlow in the Shopify App Store](https://apps.shopify.com/harperflow), open the app from Shopify admin, and sign in or link the correct HarperFlow account. Confirm the selected blog and subscription state before resuming publishing. See [the Shopify connection guide](https://www.harperflow.io/docs/connect-shopify).
If a plan change is pending or declined, review the message on HarperFlow's **Plan** page. Shopify subscriptions and plan changes require approval in Shopify, not Stripe. See [changing article volume](https://www.harperflow.io/docs/upgrading-and-downgrading).
For a support-assisted custom-app connection, contact support rather than pasting an arbitrary replacement token. Never send credentials in a public support message.
## Wix availability
Wix is in development. There is no supported self-serve Wix installation or reconnection flow to follow today. See [Wix availability](https://www.harperflow.io/docs/connect-wix).
## Still blocked?
[Contact HarperFlow](https://www.harperflow.io/contact) with the store domain, the error text, and the step that failed. Do not include passwords or access tokens.
---
---
Source: https://www.harperflow.io/docs/sitemap-diagnosis
# Sitemap diagnosis
What HarperFlow checks when it explains why a published item isn't showing up in your sitemap.
September 26, 2026
If an article looks published but isn't showing up where you'd expect, HarperFlow can check the actual item on your site and tell you exactly why. Find it on the Articles page: open a published article and click **Is it live & findable?** to run the check against your live site.
## The possible results
- **Published and sitemap-eligible**: the item is live, published, and there's nothing blocking it from appearing in your sitemap.
- **Draft**: the item exists but has never been published; drafts are never sitemap-eligible.
- **Archived**: the item was archived, which removes it from your live site and its sitemap.
- **Staged, not yet published**: the item was created or updated but hasn't had a publish step run yet, so it's sitting ready rather than live.
- **Blocked by an indexing field**: a field on the item itself, something like a "no-index" toggle, an "exclude from search" flag, or a similar indexing switch, is turned on and is telling your site to leave it out.
## What to do for each
For a draft or staged item, publish it (or wait for HarperFlow's own publish step to run, if it's a scheduled article waiting on its window). For an archived item, unarchive it if you want it live again. For an item blocked by an indexing field, check that specific field on the item. It may have been set by a platform default, a theme setting, or another integration rather than anything HarperFlow did. None of these mean the article itself was lost or needs rewriting.
> **Tip:** This check reads your site's actual live field data at the moment you run it, not a cached copy, so the result reflects reality right now rather than what things looked like the last time HarperFlow synced.
## Related
- [Publish failure error reference](https://www.harperflow.io/docs/publish-failures-error-reference)
- [Article states](https://www.harperflow.io/docs/article-states)
---
---
Source: https://www.harperflow.io/docs/structured-content-principles
# Structured content principles for GEO
The structural choices, from answer-first writing to real authorship, that make content easier for AI to retrieve.
September 26, 2026
Two articles can cover the same topic with the same underlying facts, and one gets pulled into AI answers while the other never does. Usually the difference isn't the facts. It's the structure. This chapter covers the concrete, repeatable structural choices that make content easier for a retrieval system to lift, understand, and trust.
## Lead with the answer, then explain
Retrieval systems extract passages, not entire pages. If the direct answer to a question is buried three paragraphs into a section, behind throat-clearing and context-setting, there's a real chance the system never surfaces it. The fix is straightforward, and a little uncomfortable for anyone trained to build up to a point: state the answer plainly in the first sentence or two of a section, then use the rest of the section to support, qualify, or expand on it.
This is the core idea behind Answer Engine Optimization (AEO): write each section so it could be lifted out, on its own, and still make sense as a complete answer to the question implied by its heading.
A quick before-and-after makes the difference concrete. A buried version might read: "There are a lot of factors that go into a return policy, and businesses tend to land on very different answers depending on their size and margins, but many find that giving customers more time reduces disputes and builds trust over the long run." A fronted version instead opens with: "Most retailers offer a 30-day return window," before going on to explain the tradeoffs. The second version can be lifted on its own as a complete, quotable answer. The first can't, no matter how sound the reasoning that follows it is.
## Structure for machines as well as humans
Beyond leading with the answer, the overall shape of a page matters. A few concrete habits make content dramatically easier to parse and retrieve cleanly:
- **Use real headings**, not bolded text pretending to be a heading. Headings tell both readers and retrieval systems where one self-contained idea ends and another begins.
- **Use lists and tables for anything list-shaped or comparison-shaped.** A table comparing three options is far easier for a system to extract accurately than the same comparison written as a paragraph of prose.
- **Keep paragraphs short and single-purposed.** A paragraph that tries to make three points is harder to extract cleanly than three paragraphs that each make one.
- **State facts plainly.** "The trial runs for a set number of days" retrieves worse than "the trial runs for seven days." Specificity is what makes a passage quotable.
Back this structure with real evidence wherever you're making a claim. Adding relevant statistics, citing credible sources, and including direct quotations were identified as the single highest-impact techniques for AI visibility in the foundational Princeton-led research on generative engine optimization (see [How AI search cites content](https://www.harperflow.io/docs/how-ai-search-cites-content) for the specific findings). Structure gets your content noticed; verifiable substance is what gets it trusted enough to cite.
> **Note:** None of this means writing shorter content. It means organizing longer content into more, smaller, self-contained pieces instead of one long undifferentiated block.
## Make authorship and identity unambiguous
Retrieval and ranking systems alike weigh a set of credibility signals often shortened to **E-E-A-T**: experience, expertise, authoritativeness, and trust. In practice, for most businesses, that translates into a short list of concrete things:
- A real, named author (or a clearly identified organization) attached to the content, not an anonymous byline.
- Consistent identity across the web: the same business name, description, and details wherever you appear, not conflicting versions on different platforms.
- Clear signals of who is actually behind the content and why they're a credible source on the topic.
None of this is exotic. It's the difference between content that reads as coming from somewhere specific and accountable, versus content that could have come from anywhere.
## A simple structural checklist
Before publishing a page you want an AI engine to notice, it's worth checking it against a short list:
1. Does the first sentence of each major section answer the question the heading implies?
2. Are comparisons and step-by-step processes in a list or table, not buried in prose?
3. Is every non-obvious claim backed by a number, a named source, or a quotation?
4. Is it clear, unambiguously, who wrote this and on behalf of what organization?
5. Could a single section be copied out on its own and still make complete sense?
> **Tip:** Run that checklist against a page you already have live. Most existing content fails on the first item, the answer is there, it's just three paragraphs too deep.
## How HarperFlow applies this
Every article HarperFlow publishes is built answer-first, with real headings, lists, and tables used wherever content is comparison-shaped, and a consistent, identifiable authorship and organization identity carried through automatically. That structure is enforced by the writing and rendering pipeline itself, not left to chance per article.
## Related
- [FAQ, schema, and direct answers](https://www.harperflow.io/docs/faq-schema-and-answers)
- [How AI search cites content](https://www.harperflow.io/docs/how-ai-search-cites-content)
- [Images and internal linking](https://www.harperflow.io/docs/images-and-internal-linking)
---
---
Source: https://www.harperflow.io/docs/take-a-break
# Take a break
How to pause publishing and billing for a workspace, and how it resumes.
September 26, 2026
Take a Break pauses a workspace without canceling it. Publishing stops, and for a workspace on its own Solo subscription, billing pauses too, so you aren't charged while you're not using it.
## What happens while you're on a break
Discovery, drafting, autopilot, and the due publisher all skip a workspace that's on break. If the workspace carries its own Stripe subscription, that subscription's billing is paused for the same window, with no charge in the meantime. Nothing about your site, your brand voice, or your article history changes. Everything picks back up exactly where it left off.
## Choosing a length
You can take a break for **2 or 4 weeks** at a time.
## How many you can take
You can take up to **4 breaks in any rolling 12-month window**. Only the workspace owner can start or resume one, from the Take a Break card on your Your Site page, which also shows how many breaks you have left.
## Coming back
A break ends automatically at the end of the window you chose. Publishing and billing both resume on their own, with nothing for you to do. If you're ready sooner, you can resume early from the same screen.
## Agency-managed workspaces
If your workspace is part of an Agency account, a break still pauses that workspace's publishing, but it never touches billing. Agency billing is one shared subscription across every client, and pausing it would stop billing for every other client too. See [Agency billing](https://www.harperflow.io/docs/agency-billing) for how that consolidated plan works.
> **Tip:** You'll see a banner while a break is active, plus your remaining break count, right on the Take a Break card. You don't need to remember when it ends.
## Related
- [Canceling your subscription](https://www.harperflow.io/docs/canceling-your-subscription)
- [Agency billing](https://www.harperflow.io/docs/agency-billing)
- [Reconnecting and token health](https://www.harperflow.io/docs/reconnecting-and-token-health)
---
---
Source: https://www.harperflow.io/docs/team-members-and-roles
# Team members and roles
The three account roles, what each can do, and how to invite or remove someone.
September 26, 2026
Invite teammates into your account instead of sharing your own login. Each person gets a role that controls what they can see and change across every workspace under that account.
## The three roles
- **Owner**: full control, including billing (starting a subscription, and everything covered in [Upgrading and downgrading](https://www.harperflow.io/docs/upgrading-and-downgrading) and [Canceling your subscription](https://www.harperflow.io/docs/canceling-your-subscription)). Only an owner can promote someone else to owner, and there always has to be at least one. The last owner on an account can't be removed or demoted.
- **Admin**: can invite and manage other members (except owners), and work day to day across every connected workspace. Admins can open the billing portal to switch plans, update payment details, or cancel, but can't start a brand-new subscription.
- **Member**: full working access to drafts, veto, brand voice, and the rest of the product, without member-management or billing access.
## Inviting someone
From the Members page, click **Invite teammate**, enter their email, and pick a role (it defaults to Member). HarperFlow generates a share link rather than emailing it for you. Copy it and send it however you'd like, over email, chat, or however your team normally talks. The link expires in **14 days**.
## Changing or removing access
Owners and admins can change a member's role or remove them at any time from the Members page, with one exception: the last remaining owner can't be demoted or removed, so an account can never end up without one. Removing someone takes effect immediately. They lose access across every workspace under the account right away, not just from the Members list.
> **Note:** Only an existing owner can grant the owner role to someone else. An admin can invite and manage members, but can't create a second owner.
## Related
- [Security and account deletion](https://www.harperflow.io/docs/security-and-account-deletion)
- [Accounts, organizations, and sites](https://www.harperflow.io/docs/accounts-orgs-and-sites)
---
---
Source: https://www.harperflow.io/docs/the-usage-tab
# The Usage tab
The credit-spend ledger: what shows up, what actually spends a credit, and when your count resets.
September 26, 2026
**Usage**, in your sidebar, is a read-only ledger of everything that's happened against your plan's article count this period. It doesn't let you change anything; Articles and Plan own those actions. It's there so "why did my count drop" always has a concrete answer.
## The meter at the top
A line reading how many articles you've used this period, out of your plan's total, plus the date it resets. If your plan is unmetered (some agency client workspaces have no monthly cap), it just shows how many you've used with no "of" figure, since there's nothing to divide against.
## What's in the list below it
Every event that touched your quota this period, newest first: the article's title or keyword, the date, a badge for what triggered it (**Autopilot**, **Manual**, or **Extension**), its status, and whether it used a credit.
## What actually spends a credit
An article spends a credit the moment it's created, whether autopilot wrote it, you requested it, or it came in through the browser extension. That's true regardless of what happens to it afterward: skipping a scheduled article, unscheduling it when you turn autopilot off, or taking a published one down doesn't refund the credit it already cost, since the writing itself is what you're paying for.
Runs that don't produce an article at all, because something failed partway through, or because there was nothing new left to write about, show up in the same list marked as free. They're there for visibility, not because they cost you anything.
> **Note:** A brand-new workspace with nothing written yet shows the meter with an empty list underneath, not an error. There's simply nothing to show until your first article exists.
## Related
- [Run your blog on autopilot](https://www.harperflow.io/docs/run-your-blog-on-autopilot)
- [Plans and pricing](https://www.harperflow.io/docs/plans-and-pricing)
- [Trial length and extensions](https://www.harperflow.io/docs/trial-and-extensions)
- [Article states](https://www.harperflow.io/docs/article-states)
---
---
Source: https://www.harperflow.io/docs/the-veto-digest
# The veto digest
The email HarperFlow sends after each run, listing what it just scheduled and your chance to skip anything before it publishes.
September 26, 2026
Every time a run finishes writing and scheduling new articles, whether autopilot triggered it or you did, HarperFlow emails your team a digest of exactly what it just scheduled. It's the proactive nudge that tells you it's time to look, on top of the ongoing veto window itself.
## When it's sent
The digest goes out once per completed run, right after that run's articles have their publish slots assigned. It isn't a recurring daily or weekly email; if a run produces nothing to review (for instance, every candidate topic turned out to already be covered), no digest goes out for it.
## Who gets it
Anyone on your team with the **owner**, **admin**, or **editor** role on the organization receives the digest. Viewers do not.
## What's in it
Each digest lists the articles from that run by title, alongside the exact time each is scheduled to publish, and a link back to your veto queue in the dashboard. There's nothing to click to "approve": the articles are already on the schedule. The only action available is opening the queue and skipping anything you don't want.
## The window doesn't depend on the email
The digest is a convenience, not the review mechanism itself. The actual window you have to act in runs from the moment an article is created until its own scheduled time, regardless of whether an email ever arrives. If your organization has nobody with a usable email address, or transactional email isn't configured for your account, the schedule still holds and the dashboard queue is still the source of truth.
> **Warning:** Don't treat "I didn't get a digest" as "there's nothing to review." Check your Articles queue directly if you have any doubt, especially right after a large batch.
## Related
- [How autopilot works](https://www.harperflow.io/docs/how-autopilot-works)
- [Reviewing before publish](https://www.harperflow.io/docs/reviewing-before-publish)
- [Article states](https://www.harperflow.io/docs/article-states)
---
---
Source: https://www.harperflow.io/docs/topic-clusters-and-content-types
# Topic clusters and content types
The content types and topic clusters that shape the variety of articles autopilot writes.
September 26, 2026
Two settings decide what autopilot actually writes about: content types control the format, and topic clusters control the subject matter. Together, they're what keeps a month of autopilot output from reading like the same handful of article shapes on repeat.
## Content types
HarperFlow can write in 16 formats, including comparison tables, how-to guides, buyer's guides, checklists, case studies, decision guides, and troubleshooting guides. Onboarding starts you with all 16 in play; you can narrow the list to just the formats that fit your site from your project settings at any time, and broaden it again later if you change your mind. HarperFlow picks whichever allowed format best fits a given keyword itself, rather than you choosing one per article.
## Topic clusters
Topic clusters are the subject areas your content covers, drawn from your niche during onboarding. HarperFlow drafts an initial set of 12 candidates for you to review, and you can edit that list, renaming, removing, or adding to it, into a final reviewed set of anywhere from 12 to 24 clusters before autopilot is allowed to run.
## Why the minimum matters
Autopilot requires at least 12 reviewed, active clusters before it will dispatch a run. This isn't a soft recommendation: a project sitting below that line simply won't be picked up until you're back above it, no matter how healthy the rest of its setup is.
## Resetting onboarding
If you reset onboarding and go through cluster review again, your previous clusters aren't deleted. They're marked inactive instead, so any already-published articles keep their original topic-cluster association even though the cluster itself no longer counts toward your active set.
> **Warning:** A big edit to your topic clusters, or a reset, can temporarily drop you below the 12-cluster minimum and pause autopilot until you've reviewed enough of them again.
## Related
- [Org Brain and shared memory](https://www.harperflow.io/docs/org-brain-and-shared-memory)
- [How autopilot works](https://www.harperflow.io/docs/how-autopilot-works)
- [The onboarding wizard, step by step](https://www.harperflow.io/docs/onboarding-wizard)
---
---
Source: https://www.harperflow.io/docs/topic-runs
# Topic runs
How to seed a single run with your own topic instead of leaving keyword discovery to HarperFlow.
September 26, 2026
Most runs start with HarperFlow discovering a keyword on your behalf. You can also hand it a specific topic yourself and get exactly one article written on that subject, without waiting for discovery to surface it. You'll find this option alongside the default "let HarperFlow discover a keyword" choice, wherever you start an on-demand run.
## What you can type
A topic needs to be a plain phrase, not a URL: no `http://` links and no `www.` addresses. It has to land between 3 and 100 characters, using letters, numbers, spaces, apostrophes, and hyphens only. Anything else comes back with a plain-language reason instead of a cryptic error, so you can fix it and try again.
## What happens to it
HarperFlow cleans up whatever you type into a search-shaped phrase (lowercase, no stray punctuation) and checks it against real keyword volume data, so the article is still built around a term people actually search for, not just your exact wording. That one topic becomes the only candidate for the run: there's no separate discovery step, no niche scoring, and no competing keyword that can win out over the one you asked for. The result reads like any other HarperFlow article: full body, citations, images, and metadata, just built around your exact subject instead of a discovered one.
## Duplicate protection
Before the run starts, HarperFlow checks whether that topic, or something close enough to it, is already published or mid-flight for this project. If it is, you're told so directly instead of quietly getting a near-duplicate article. If you deliberately want to cover the same ground again anyway, for instance refreshing an older take, the request can force past that check instead of being blocked by it.
> **Tip:** An "already covered" result isn't a dead end. It usually points at an existing article worth looking at, updating, or linking from, rather than one you need HarperFlow to rewrite from scratch.
## Related
- [On-demand runs](https://www.harperflow.io/docs/on-demand-runs)
- [Org Brain and shared memory](https://www.harperflow.io/docs/org-brain-and-shared-memory)
- [Article states](https://www.harperflow.io/docs/article-states)
---
---
Source: https://www.harperflow.io/docs/topical-authority-and-clusters
# Topical authority and content clusters
Why depth across many connected articles, not single posts, builds the topical authority AI engines recognize.
September 26, 2026
A single excellent article rarely earns lasting AI visibility on its own. What earns it is depth: a body of connected content that thoroughly covers a subject from multiple angles, kept current over time. This chapter explains why depth and freshness function as a compounding pair, and how to think about building both deliberately.
## Why one article isn't enough
A retrieval system doesn't just look for the single best answer to one question. Over time, and across many different but related queries, it's effectively sampling how thoroughly a source covers a subject. A site with one solid article on a topic can answer one question well. A site with fifteen connected articles, each covering a distinct angle of the same subject, can answer fifteen related questions well, and each one reinforces that the site genuinely knows the subject rather than having gotten lucky once.
That's the practical meaning of **topical authority**: not a score assigned to you by any single engine, but the accumulated effect of covering a subject thoroughly enough that you become a recognizable, repeated source across many of the questions in that space.
## Depth requires more content than most teams expect
This is the part that surprises people: real topical depth requires genuine volume, not a handful of cornerstone posts. Covering a subject's real breadth, definitions, comparisons, common mistakes, edge cases, how-to specifics, current developments, naturally produces many articles, not few. That's simply a function of how many distinct, real questions exist inside any subject worth being known for.
This is also, honestly, why most GEO efforts stall before they compound: the volume required to build genuine depth is more than most teams can sustain by hand, on top of everything else they're already doing.
## Clusters: the practical unit of depth
Rather than planning topics one article at a time, it helps to plan in **clusters**: a group of related articles that together cover one subject area thoroughly, structured so each article can stand alone but also clearly belongs to the same body of work. A cluster might cover a single product category, a single stage of a buyer's journey, or a single recurring customer question along with its variations.
Planning this way does two things a random list of blog ideas doesn't. It surfaces gaps deliberately (which obvious sub-questions in this subject haven't we answered yet), and it avoids accidental duplication (have we already covered this angle under a different title).
As a concrete shape: a cluster built around, say, employee scheduling might include one article defining the problem, one comparing common approaches, one walking through a specific method step by step, one covering frequent mistakes, and one addressing a narrow edge case that keeps coming up in customer questions. None of those five articles substitutes for another. Together, they cover the subject the way a single 3,000-word "ultimate guide" rarely does in practice, because a retrieval system still has to extract a self-contained passage from wherever the answer lives.
> **Note:** A cluster with strong internal linking between its own articles also reinforces retrievability directly. It gives both readers and retrieval systems a clear signal that these pieces belong to one coherent body of knowledge on the subject.
## Freshness is not optional
Depth alone doesn't hold. Content that goes stale loses ground. According to GEO Compass's analysis of structured data and freshness signals for answer and generative engine optimization, pages left unrefreshed lose AI citations at roughly three times the rate of pages that are kept current. Recency functions as a trust signal in its own right, independent of how good the original content was when it was published.
In practice this means treating your existing library as an asset to maintain, not just a backlog to add to. A quarterly pass through your highest-value articles, updating figures, examples, and any claims that have aged, protects citations you already earned rather than only chasing new ones.
> **Tip:** When deciding what to update first, start with your highest-traffic or highest-intent articles, not your newest ones. Losing citations on the pages doing the most work costs more than a stale page nobody was finding anyway.
## Consistency compounds the effect
Depth and freshness both reward the same underlying behavior: publishing on a real, sustained cadence rather than in bursts. Each new, well-structured article adds retrievable passages answering new questions, and a library that's visibly still growing and still being maintained reads as more authoritative than one that went quiet eighteen months ago. The effect is cumulative: authority built this way tends to compound month over month rather than plateau.
## How HarperFlow applies this
HarperFlow plans content in reviewed topic clusters rather than one-off ideas, tracks which clusters still have coverage gaps, and lets you generate a new article directly against a specific gap. Its Brain view shows cluster coverage and article counts at a glance, and shared memory keeps that coverage picture unified across every site in an organization that chooses to share it, so depth is planned deliberately rather than accumulated by accident.
## Related
- [Topic clusters and content types](https://www.harperflow.io/docs/topic-clusters-and-content-types)
- [Org Brain and shared memory](https://www.harperflow.io/docs/org-brain-and-shared-memory)
- [How autopilot works](https://www.harperflow.io/docs/how-autopilot-works)
- [Measuring GEO success](https://www.harperflow.io/docs/measuring-geo-success)
---
---
Source: https://www.harperflow.io/docs/trial-and-extensions
# Trial length and extensions
What happens when your trial ends, and how to get more time if you need it.
September 26, 2026
[Your free trial](https://www.harperflow.io/docs/your-free-trial) covers what the 7-day, 3-article trial includes and how eligibility works. This article covers what happens at the end of it, and how extensions actually work, since neither is automatic.
## What happens when it ends
Your trial converts automatically into whichever paid plan you selected when you started it. There's no separate "activate" step: if you don't cancel first, the card on file is charged for that plan the moment the trial period ends. If you do want to stop before then, cancel from the billing portal. See [Canceling your subscription](https://www.harperflow.io/docs/canceling-your-subscription).
## Getting more time
If you need longer than 7 days or more than 3 articles to decide, several places in the app (the plan screen while you're trialing, the setup gate, the Enterprise card) offer a short call with HarperFlow's founder instead of a self-serve button. This is a real, manual conversation, not an automatic top-up: whoever takes the call decides how much extra time or how many extra articles make sense for your situation, based on what you tell them.
## Promo codes
If you were given a promotional code, from the team, a partner, or a specific offer, redeem it from the "Have a code?" box on your Plan page. Codes can grant bonus articles, extra trial days, or occasionally a free month, depending on what was issued. They aren't something you request yourself; they come from the team or a partner program.
> **Warning:** Starting a trial requires payment details upfront, even though you aren't charged during the 7 days. If you don't cancel before the trial ends, you will be charged for the plan you chose at signup. There's no separate reminder step that pauses this for you.
## Related
- [Your free trial](https://www.harperflow.io/docs/your-free-trial)
- [Plans and pricing](https://www.harperflow.io/docs/plans-and-pricing)
- [Canceling your subscription](https://www.harperflow.io/docs/canceling-your-subscription)
---
---
Source: https://www.harperflow.io/docs/unpublish-and-take-down
# Unpublish and take-down
How to pull a live article back off your site, and how to bring it back later if you change your mind.
September 26, 2026
Sometimes an article that already went live turns out to be wrong for your site after all. **Take this down** reverts it without losing the work behind it.
## What it does
Taking an article down reverts the live item on your connected site back to a draft (on Webflow, this uses the same mechanism as unpublishing an item there directly). The content stays intact in your CMS and in HarperFlow; only its public visibility changes. It's the recoverable alternative to deleting something outright.
People typically reach for this after spotting a factual mistake once an article is already live, after a product change makes something outdated, or simply because a piece reads differently once it's actually on the site than it did in review.
## It's safe to click more than once
If an article is already off your site, for instance because you unpublished it directly in your CMS, clicking **Take this down** in HarperFlow simply confirms that state rather than failing or erroring.
## Bringing it back
A taken-down article shows a **Restore** option, the same one skipped articles get. Restoring sends it back to pending review with its old publish time cleared, so you'll need to give it a fresh slot; see [Article states](https://www.harperflow.io/docs/article-states).
## Your quota isn't affected either way
Taking an article down, or restoring one afterward, doesn't touch your article quota in either direction. That was already spent the moment the article was first written, so pulling it offline temporarily doesn't refund it, and bringing it back doesn't spend a new one.
## Who can do this
Taking articles down, and restoring them afterward, is limited to your organization's owners and admins.
> **Tip:** Use Take this down for anything you need off your site right away and might want back later. It's built to be reversible, so there's little downside to using it as soon as you spot a problem.
## Related
- [Article states](https://www.harperflow.io/docs/article-states)
- [Reviewing before publish](https://www.harperflow.io/docs/reviewing-before-publish)
- [Quality and trust panel](https://www.harperflow.io/docs/quality-and-trust-panel)
---
---
Source: https://www.harperflow.io/docs/upgrading-and-downgrading
# Changing your article volume
Change your monthly volume through the web billing flow or Shopify approval screen.
September 26, 2026
Open HarperFlow's **Plan** page and choose the monthly article volume that matches your publishing cadence. The available single-site range is 5–30 articles. The full publishing pipeline is included at every volume.
## Webflow and WordPress
Use the volume controls for your current web subscription and review the billing confirmation before accepting the change. Web subscriptions use Stripe. Check the amount, any adjustment, and the effective date displayed in the confirmation flow. The payment-method portal and cancellation control are separate actions.
## Shopify
Open HarperFlow from **Apps** in Shopify admin, then open **Plan**. Select your new article volume and continue to Shopify's subscription approval screen. Review and approve the replacement recurring charge there; the change is not completed merely by moving the volume control.
Shopify calculates applicable billing adjustments. The confirmation screen is the authority for the charge and effective timing. Your Shopify subscription uses the [Shopify price ladder](https://www.harperflow.io/docs/plans-and-pricing), not the Webflow/WordPress formula.
## Extra articles
If you need articles beyond your current monthly volume, use the extra-article control shown for your subscription. Shopify add-ons require their own Shopify charge approval. Check the quantity and total before confirming.
## After confirming
Return to HarperFlow and check the active volume on **Plan**. If the page shows a pending or declined change, follow its instructions or [contact support](https://www.harperflow.io/contact). Do not start another subscription to work around an unresolved update.
---
---
Source: https://www.harperflow.io/docs/webflow-connection-issues
# Webflow connection issues
Common reasons a Webflow connection breaks, and how the token setup avoids them.
September 26, 2026
Webflow is HarperFlow's most complete integration, and most connection problems trace back to the site API token it runs on.
## Common causes
- **The token was regenerated or revoked.** Generating a new token in Webflow's Site settings, or revoking the old one, invalidates whatever HarperFlow was using. HarperFlow only ever holds the one token you gave it. It has no way to know a new one exists until you paste it in.
- **The token's permissions were narrowed.** If the token no longer has read/write access to Sites, CMS, Assets, Pages, Custom code, and Components, calls that need those permissions start failing even though the token itself is still technically valid.
- **The site moved workspaces, or was deleted.** A token is scoped to whatever it could see at the time it was generated; moving the site to a different Webflow Workspace or deleting it removes what HarperFlow was authorized to reach.
Any of these show up the same way: a reconnect banner on your dashboard, and publishing paused until it's fixed. See [Reconnecting and token health](https://www.harperflow.io/docs/reconnecting-and-token-health) for what that banner means and the 14-day timeline behind it.
## "This token cannot see any sites"
This specific message means the token you pasted is scoped to your whole Webflow Workspace rather than to one site. Generate a fresh token scoped to the individual site instead. See [Connect your Webflow site](https://www.harperflow.io/docs/connect-webflow) for exactly which permissions to turn on.
## Fixing it
Reconnecting is almost always faster than trying to repair the old token: generate a brand-new one with the same permissions as the original setup, and paste it into the reconnect flow on your Your Site page.
> **Tip:** If you're not sure whether a token still works, generating a fresh one and reconnecting is the fastest way to rule it out. There's no need to diagnose the old one first.
## Related
- [Reconnecting and token health](https://www.harperflow.io/docs/reconnecting-and-token-health)
- [Connect your Webflow site](https://www.harperflow.io/docs/connect-webflow)
- [Custom field mapping](https://www.harperflow.io/docs/custom-field-mapping)
---
---
Source: https://www.harperflow.io/docs/what-harperflow-does
# What HarperFlow does
HarperFlow researches, writes, schedules, and publishes articles to Webflow, WordPress, and Shopify.
September 26, 2026
HarperFlow is a publishing pipeline for **Webflow, WordPress, and Shopify**. It researches topics, writes source-backed articles in your brand voice, schedules them, and publishes to your connected blog. Articles are built to be cited by AI; rankings, traffic, and citations are not guaranteed. **Wix is in development and is not available for self-serve publishing.**
## The autopilot cycle
1. **Discover:** identify questions and topics that fit your business and audience.
2. **Draft:** research and write a complete article with cited sources, structure, images, and metadata.
3. **Schedule:** assign a publishing slot so articles arrive on a steady cadence.
4. **Veto:** review the email digest and skip any scheduled article you do not want published.
5. **Publish:** articles you have not vetoed publish automatically at their scheduled time.
There is no mandatory approval step for every article. On-demand runs also create scheduled drafts; check their publishing time and use the available skip controls before they go live.
## Connect your platform
- [Webflow](https://www.harperflow.io/docs/connect-webflow): connect your site and map or provision a CMS collection.
- [WordPress](https://www.harperflow.io/docs/connect-wordpress): connect your site using the supported WordPress setup.
- [Shopify](https://www.harperflow.io/docs/connect-shopify): install [HarperFlow from the Shopify App Store](https://apps.shopify.com/harperflow), choose a blog, and approve billing through Shopify.
The publishing engine is shared, but CMS field mapping, theme rendering, and billing differ by platform. See [pricing](https://www.harperflow.io/docs/plans-and-pricing) and [the setup wizard](https://www.harperflow.io/docs/onboarding-wizard).
---
---
Source: https://www.harperflow.io/docs/what-is-geo
# What is GEO?
A plain-language definition of Generative Engine Optimization and why it matters alongside traditional SEO.
September 26, 2026
If you've heard the term GEO in the last year or two, you're not behind. It's a genuinely new discipline, not an old idea with a new name. This chapter gives you a working definition, distinguishes it from the terms it gets confused with, and explains why it's worth learning now rather than later.
## The one-line definition
**Generative Engine Optimization (GEO)** is the practice of structuring content so that generative AI systems, tools like ChatGPT, Perplexity, Google's AI answers, and Claude, can find it, trust it, and cite it in the answers they generate. Where search engine optimization aims to rank a page in a results list, GEO aims to get a business named and linked inside an AI-written answer.
The term comes from a specific piece of research: a 2024 study led by researchers at Princeton, published at the ACM KDD conference, that tested content-optimization techniques against real generative engines and measured which ones actually moved the needle on visibility. That paper is why GEO has a specific, testable meaning rather than being marketing shorthand.
## GEO, AEO, and SEO: three related but distinct things
These three acronyms get used almost interchangeably, which causes real confusion. Here's how to keep them straight:
- **SEO (Search Engine Optimization)** is the familiar discipline of ranking web pages in a traditional search engine's results list, built around keywords, backlinks, and page authority.
- **AEO (Answer Engine Optimization)** is about shaping a piece of content so it directly answers the question someone asked, making it the clean, quotable response an engine reaches for. AEO is as much about format, a clear answer, stated plainly, near the top, as it is about topic.
- **GEO (Generative Engine Optimization)** is the broader discipline of getting cited inside a generative AI's synthesized answer at all, whether that answer quotes you directly, summarizes you, or links to you as a source.
In practice, the three overlap heavily, and the working assumption in this guide (and across most of the industry) is that they describe the same underlying goal from different angles: being the source an AI system chooses to trust. GEO does not replace SEO. It extends it into a world where the "result" a person sees is a generated answer, not a list of ten blue links.
## Why this isn't just SEO with new branding
Classic SEO ranking rewards things like backlink volume, keyword density, and domain age, signals a crawler can measure at scale. Generative engines work differently. Most of them use a technique called retrieval-augmented generation (RAG): before writing an answer, the system retrieves a handful of relevant documents, then generates a response grounded in what it retrieved, usually with citations attached. That retrieval step rewards different properties than a keyword-matching crawler does: whether a passage plainly states what it means, whether it can be lifted cleanly out of the page, and whether independent sources agree on the same facts.
That's a meaningfully different target to write for, which is why GEO earned its own name instead of becoming a subsection of an SEO playbook.
> **Note:** If you want the mechanics of how engines actually choose what to retrieve and cite, that's covered in detail in [How AI search cites content](https://www.harperflow.io/docs/how-ai-search-cites-content). This chapter stays at the definitional level on purpose.
## Why it matters now, not eventually
The practical reason to learn GEO today rather than in a year is timing. ChatGPT alone has grown to roughly one billion monthly users, according to DemandSage's usage statistics, which makes it a genuine new front door for how people discover businesses, not a niche experiment. The shift isn't limited to consumer curiosity, either: in a multi-source analysis of B2B buying behavior distributed via PR Newswire, 94% of B2B buyers said they now use AI tools somewhere in their purchase research.
Put simply, a large and fast-growing share of the people who might become your customers are asking an AI about your category before they ever type a query into a traditional search box. If your content isn't structured in a way that AI systems can retrieve and trust, you aren't ranked poorly in that conversation. You're simply not part of it.
## What this means in practice
None of this requires abandoning what you already know about writing good content. It means adding a layer on top: making sure claims are backed by real sources, making sure an answer is stated plainly instead of buried under three paragraphs of preamble, and making sure the underlying structure, headings, lists, clear sections, is something a machine can parse as cleanly as a person can. The rest of this guide walks through each of those pieces in more depth.
> **Tip:** A fast way to see where you stand today: ask ChatGPT, Perplexity, and Google's AI answers a question in your own category and see whether your business shows up, and whether what they say about you is accurate. That single exercise tells you more than any theoretical framework.
## Related
- [How AI search cites content](https://www.harperflow.io/docs/how-ai-search-cites-content)
- [GEO vs. SEO: what's actually different](https://www.harperflow.io/docs/geo-vs-seo)
- [What HarperFlow does](https://www.harperflow.io/docs/what-harperflow-does)
---
---
Source: https://www.harperflow.io/docs/wordpress-connection-issues
# WordPress connection issues
Common reasons a WordPress connection breaks, from Application Passwords to REST access.
September 26, 2026
WordPress connects over HTTPS using an Application Password rather than a login you share, so most connection problems trace back to that credential or to the account it belongs to.
## Common causes
- **The Application Password was revoked or regenerated.** WordPress lets you revoke individual Application Passwords from Users → Profile without affecting your main login. If the one HarperFlow uses gets revoked or a new one is generated in its place, the old one stops working immediately.
- **The connected user account changed.** If the WordPress user HarperFlow authenticates as is deleted, or demoted below Editor/Administrator, HarperFlow loses the permissions it needs to publish.
- **A security plugin or host firewall started blocking REST requests.** Some security plugins and hosting firewalls can restrict or disable the WordPress REST API ( `/wp-json/`) after the fact, which is what HarperFlow publishes through.
- **The site lost HTTPS.** Application Passwords require HTTPS; a lapsed or misconfigured SSL certificate breaks the connection until it's restored.
## What you'll see
A broken WordPress connection shows the same reconnect banner as any other platform, and publishing pauses until it's fixed. See [Reconnecting and token health](https://www.harperflow.io/docs/reconnecting-and-token-health) for the general flow and the 14-day timeline before HarperFlow pauses billing on an unresolved connection.
## The SSO key
If one-click sign-in from your wp-admin stops working specifically (while publishing still works, or vice versa), it's usually the separate SSO key rather than the Application Password. Reconnecting resets both.
## Fixing it
Reconnect from your Your Site page using the same three fields as the original setup: your site URL, your WordPress username, and a freshly generated Application Password. See [Connect your WordPress site](https://www.harperflow.io/docs/connect-wordpress) for exactly where to generate one.
> **Note:** Reconnecting doesn't change where your articles go or how they're categorized. It only refreshes the credential HarperFlow authenticates with.
## Related
- [Reconnecting and token health](https://www.harperflow.io/docs/reconnecting-and-token-health)
- [Connect your WordPress site](https://www.harperflow.io/docs/connect-wordpress)
---
---
Source: https://www.harperflow.io/docs/your-free-trial
# Your free trial
Understand the Webflow/WordPress trial and the separate Shopify App Store trial.
September 26, 2026
HarperFlow offers a 7-day trial. The signup and billing flow depends on the platform you connect.
## Webflow and WordPress
The web trial includes **3 articles over 7 days**. Payment details are required at checkout. Review the selected volume, renewal amount, and trial end date before starting. Cancel through HarperFlow's **Plan** page before the trial ends if you do not want the subscription to renew.
## Shopify
Install [HarperFlow from the Shopify App Store](https://apps.shopify.com/harperflow). The listing offers a **7-day free trial**, with subscriptions billed through Shopify from **$39.90/month**. Review the trial terms and recurring charge on Shopify's approval screen. The web trial's Stripe checkout is not the signup path for a Shopify App Store subscription.
## Scheduled publishing during evaluation
Check your generated articles and their scheduled times in HarperFlow. The veto digest lets you skip articles before they publish. A trial is not a guarantee that every draft will wait indefinitely for manual approval.
## Eligibility and more time
Trial eligibility is checked during signup. A new account, reinstallation, or reconnecting the same site does not guarantee another trial. If you need help completing evaluation, [contact the founder](https://www.harperflow.io/contact); an extension is not automatic, and any Shopify billing adjustment must be confirmed for that subscription.
See [connect Shopify](https://www.harperflow.io/docs/connect-shopify), [pricing](https://www.harperflow.io/docs/plans-and-pricing), and [cancellation](https://www.harperflow.io/docs/canceling-your-subscription).
---