Tool

AI visibility check

See whether ChatGPT, Claude and Perplexity can reach your site, read it and understand it. 20 checks, a score out of 100, and a fix for every check that does not pass.

No sign-up. Results in seconds. I check one page: the one at the address you enter.

Method

How the score works

A passed check earns its full points, a partly met one half, a failed one zero. The weights are uneven on purpose: without crawler access and content in the HTML, nothing else matters.

Access30 pts

Whether AI crawlers can reach the page at all.

AI crawlers are allowed 12 pts
I check robots.txt for the crawlers that fetch answers live: OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User and PerplexityBot. Blocking training crawlers such as GPTBot is your call and costs no points.
No indexing block 8 pts
A noindex meta tag or X-Robots-Tag header removes the page from search engines, and with them from AI answers.
Sitemap listed in robots.txt 6 pts
A sitemap referenced in robots.txt lets crawlers find every page, including the poorly linked ones.
llms.txt file 4 pts
llms.txt is a short index of your key content for language models. Good practice, but the major models have not confirmed they use it, so it weighs little.

Readability25 pts

Whether the content is in the HTML and clearly structured.

Content in raw HTML 13 pts
Most AI crawlers do not run JavaScript. If content only appears after scripts load, the crawler sees an empty page.
Headings form an outline 7 pts
H2 and H3 headings without skipped levels split content into chunks a model can quote on their own.
One H1 heading 5 pts
A single H1 states what the page is about. None, or several, blurs the topic.

Structured data20 pts

Whether a machine can tell who is behind the page and what it is about.

Who is behind the site 6 pts
Organization or Person schema with sameAs links (LinkedIn, company profiles) lets AI connect the site to your brand.
Valid JSON-LD 5 pts
JSON-LD structured data describes the page in machine terms. One syntax error voids the whole block.
Page type in schema 4 pts
Article, Service, Product or BreadcrumbList tell what this specific page is and where it sits in the site.
Meta description 3 pts
A 70–160 character meta description is a ready summary of the page for search engines and AI.
Open Graph tags 2 pts
og:title, og:description and og:image decide how the page looks when shared or previewed by an assistant.

Trust12 pts

Signals of freshness, authorship and one clear address.

Machine-readable dates 4 pts
Models prefer current sources. Published and updated dates in schema or a time element show the content is fresh.
Author and contact 3 pts
A link to an about or contact page and an author in schema signal that a real person or company stands behind the content.
Canonical URL 2 pts
A canonical link points to one correct address, so citations do not scatter across duplicates.
HTTPS 2 pts
An encrypted connection is the baseline of trust for search engines and browsers.
Declared language 1 pts
The lang attribute tells which language the content is in, so it reaches answers in that language.

Citability13 pts

Whether the content answers the questions people ask AI.

Question-style headings 5 pts
People ask AI in full sentences. A question heading with the answer right below it is the easiest chunk to quote.
FAQ section 4 pts
A questions and answers section gathers what customers ask most in one place.
FAQPage schema 4 pts
FAQPage describes questions and answers in JSON-LD, so a machine does not have to guess them from the layout.

What the check does not measure

The score tells you whether AI can reach, read and understand the page. It does not tell you whether AI will cite you. That also depends on whether your content answers real customer questions, and on mentions of your brand elsewhere. I cover the full picture in the guide to AI search visibility.

To see how B2B tech and SaaS companies score, read my study of 46 websites.

FAQ

About the check.
Short answers.

How the check works and what to do with the score.

I usually reply the same business day.

What exactly does the check test?

I fetch the page the way an AI crawler does, along with robots.txt, llms.txt and the sitemap. I check 20 things in five groups: crawler access, readability, structured data, trust signals and citability. The full list with weights is in "How the score works".

Does a high score mean ChatGPT will cite my site?

No. A high score means nothing technical stands in the way. Citations also depend on content that answers questions directly, and on brand authority, meaning mentions of you in other sources.

Why does the check read raw HTML instead of the rendered page?

Because that is what most AI crawlers do. GPTBot, ClaudeBot and PerplexityBot do not run JavaScript. If content only appears after scripts load, the crawler sees an empty page, even if everything looks fine in a browser.

Should I block AI crawlers in robots.txt?

Let in the search crawlers that fetch pages live (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot), because they decide whether you show up in answers. You can block training crawlers such as GPTBot with no effect on this score.

What happens to my URL and email?

The URL is only used to fetch its public files once. The result stays in server memory for 15 minutes so the report does not fetch the site again. You only give your email if you want the fixes: I send the report and save the contact so I can reply if you write back. No newsletter.

Contact

Got a project?
Get in touch.

Tell me in a few sentences what you want to build and by when.

  1. You write what you want to build
  2. I reply within 24h
  3. We agree on scope and price

What do you need
Project budget

By sending this form you agree to have your data processed to reply to your message. Details in privacy policy.