Free Chrome extension · No account · Runs in your browser

Can AI assistants
read your website?

FlowPro checks whether ChatGPT, Claude, Perplexity and Gemini can access, read and reference your site — and fixes what they can't. Start with the free Chrome extension. The audit costs nothing; the fixes are priced to what you need.

17 checks In four groups
0–100 Readiness score
Free Audit and fix list
By estimate Work priced to the scope
The problem

Four things that stop AI assistants from reading a site

🚫

Crawlers blocked by accident

A single Disallow written for one bot, or a firewall rule, keeps GPTBot, ClaudeBot or PerplexityBot out. OpenAI, Anthropic and Perplexity each document this: a blocked search bot means your pages are not in the pool of sources.

📄

Files that lie

An llms.txt or sitemap.xml returns status 200 but serves the home page as HTML. Status codes pass; content fails. Our own site did this before we checked.

🏷

No machine-readable facts

Company name, services, prices and FAQ exist only as design, not as visible text or JSON-LD. Systems that fetch a page read the HTML, not the layout.

✂️

Slogan first, answer last

The opening paragraph says “Small agency. Big-team pace.” and the useful sentence is in section four. Studies of LLM citations find the first third of a page is where most cited text comes from.

How it works

Three steps from “not sure” to fixed

1

Install and run the check

Open any page of your site, click the extension. It reads the page and the site's public files (robots.txt, sitemap.xml, llms.txt) in your browser and shows a 0–100 score with one of four verdicts: Hard for AI to read, Partly readable by AI, Readable by AI, Ready to be cited.

Free
2

Read the prioritised fix list

Every failed check comes with what it is, why it matters and what to change. Blockers first, improvements second. Many fixes take a developer ten minutes.

3

Send it to me, or do it yourself

No developer? Send me the address. I go through your pages myself, tell you which fixes actually matter and what they involve, and send an estimate for the ones you want. See how the work is scoped →

Audit and estimate free
The 17 checks

What the 17 checks cover

Rules come from OpenAI, Anthropic, Perplexity and Google documentation and from published studies. The weight of each check follows the strength of its evidence.

🤖

Crawlability

  • robots.txt rules for the AI crawlers of OpenAI, Anthropic, Perplexity, Google, Bing, Common Crawl and Apple
  • sitemap.xml present and valid
  • llms.txt present and served as text
{ }

Structured data

  • JSON-LD found and parseable
  • Organization, Service or Product, FAQPage where the page content justifies it
  • No contradictions with visible text
M

Meta tags

  • title 30–60 characters
  • description 70–160 characters
  • canonical
  • Open Graph complete
  • Language declared
H

Content structure

  • One H1
  • Heading hierarchy
  • Opening paragraph under 60 words that states what the page is about
  • Q&A block with direct answers
What the score is — and isn't

It measures whether your site is accessible and readable for AI systems, using rules from platform documentation and published studies. It does not measure whether AI assistants mention you: most brand mentions in AI answers come from third-party pages, not from the brand's own site. The score tells you your side of the job is done.

Done for you

No developer? I'll implement the fixes

You decide which of these are worth doing. I scope them with you and price the estimate to that, so you are not paying for a package with parts you don't need.

  • robots.txt rules for every documented AI search and user bot, plus a written decision on training bots
  • sitemap.xml and llms.txt created or repaired and verified by content-type, not just status code
  • JSON-LD (Organization, Service/Product, FAQPage) on up to 5 key pages
  • Titles, descriptions, canonical and Open Graph on the same pages
  • Opening paragraphs and heading hierarchy rewritten with you; FAQ block added where it belongs
  • Re-check with the extension: every fixed page reaches Ready to be cited; 30-day follow-up re-check included
What it costs
Free
audit, fix list and estimate
The work itself is quoted to the scope you agree. Most small-business sites land between 2 000 and 6 000 zł net; you see the number before anything starts.
Invoiced in PLN, or in EUR for clients outside Poland. Not VAT-registered (art. 113 VAT Act) — the net figure in the estimate is the final figure.
Who does the work

One engineer, start to finish

You work directly with the person who reads your site, writes the files and edits the templates. Nothing is passed to a subcontractor, and nothing waits behind an account manager.

🎯
You talk to the person who reads your site, writes the files and edits the templates. No hand-offs.
📋
Everything I change is listed in a report you can hand to your next developer.
🔒
Access is scoped and revoked at the end. The audit itself never leaves your browser.
I
Ivan Karabeinikau
Founder & Senior Engineer · FlowPro
“Ten years of building web platforms, most recently as tech lead for a bank's frontline platform. I built the checker because my own site failed half of it. Now I fix the same list for other people.”
More about me →
!
What I don't promise

I don't promise that ChatGPT, Claude, Perplexity or Gemini will mention or recommend you. Nobody can: the platforms decide, and most brand mentions come from other people's pages. What I promise is narrower and checkable: every fixed page passes the same free check you can run yourself, and the files behind it say the right thing when a crawler asks.

FAQ

Questions people ask before installing

Is the extension really free?

Yes. It runs entirely in your browser, makes no account and sends nothing to us. So is the audit I do myself if you send me your address. Only the implementation work is paid, to an estimate you agree first.

What does “Ready to be cited” mean?

It is the top verdict of the checker: crawlers are allowed, files are valid, facts are in visible HTML and JSON-LD, and every page starts with a direct answer. It means the technical preconditions are met. It is not a promise that any assistant will cite the page.

Do I need llms.txt?

For Google Search, no — Google states it ignores the file. No AI platform has confirmed reading it, and a 2026 study of 137 000 sites found 97 % of llms.txt files get no requests. Coding agents do read it. The checker treats it as a small, cheap improvement, not a blocker.

Does structured data get me cited?

Adding JSON-LD did not increase AI citations in the largest controlled study so far (Ahrefs, 1 885 pages, 2026). It still helps search engines disambiguate your business and qualifies pages for rich results. The checker scores it as useful hygiene, weighted below crawler access and content structure.

What if my site is on a custom platform?

Send the URL through the form. I look at what the fixes would actually involve on your stack and say so within one business day. If something is impractical there, I tell you that instead of quoting it.

Do you still do automation for real-estate agencies?

Yes, as a separate offer at flowpro.dev/automatyzacja (Polish).

Check your site now. It takes ten seconds.

Install the extension, open your home page, click the icon. If the verdict is anything but Ready to be cited, you'll know exactly what to fix — and who can do it.

Install the free extension
or write directly: ivan@flowpro.dev