Free · No login · 22 checks

Free AI visibility checker: can ChatGPT, Claude and Perplexity read your website?

We fetch your page the way AI crawlers do, without JavaScript, and show what stops them reaching and reading it.

Any public page works: your homepage, a blog post, a product page.

Up to 20 secondsWe never run your JavaScriptHow the check works

Example: what an AI crawler receives, no JavaScript
  • <html lang="en">
  • <title>Your Business</title>title
  • <meta name="description">description
  • <script type="application/ld+json">Organization only
  • <link rel="canonical">canonical
  • <body>
  • <div id="root"></div>0 words
  • <script src="/app.js"></script>
A page built in the browser can look fine to you and arrive at most AI crawlers as a title and nothing else. That is the kind of thing the check finds.

What you get

  • A score out of 100 and a grade
  • The three fixes worth the most points
  • 22 checks, each with what we found and how to fix it
  • What your robots.txt says to 22 AI crawlers
  • A prompt to hand your developer, and a markdown copy

It checks whether AI systems can reach and read your site. It doesn't track whether they mention you.

Learn

What AI crawlers need from your site

Before an AI assistant can quote your page, it has to be allowed in, find the page, read it and make sense of it. The check scores those four things, 22 checks in all, out of 100.

Google renders JavaScript. OpenAI, Anthropic and Perplexity read the HTML your server sends. The check is built for the second group.

  1. 25 points

    Let them in

    Your robots.txt decides which crawlers may fetch your pages. One stray Disallow line, or a noindex tag, keeps AI tools out without anyone noticing.

    We look at:

    • robots.txt
    • rules for each AI crawler
    • blanket blocks
    • noindex
  2. 25 points

    Help them find your pages

    A sitemap lists every page, and a Sitemap line in robots.txt says where it is. A canonical link names the one true address for each page. A feed lists what's new.

    We look at:

    • XML sitemap
    • robots.txt Sitemap line
    • canonical URL
    • RSS or Atom feed
    • llms.txt
  3. 30 points

    Make it easy to read

    AI systems split a page up by its headings and lean on its title, description and structured data to know what it is about, who wrote it and when.

    We look at:

    • title and description
    • one H1
    • JSON-LD
    • Open Graph
    • landmarks
    • language
    • dates
    • a markdown copy
  4. 20 points

    Don't hide it behind JavaScript

    Most AI crawlers, including OpenAI's, Anthropic's and Perplexity's, read the HTML your server sends and don't run JavaScript. Google does render JavaScript. A page drawn in the browser can look blank to the rest.

    We look at:

    • text in the raw HTML
    • app shells
    • head tags before rendering
    • page weight
    • bot challenges
    • redirects
About llms.txt. It's a proposal (llmstxt.org (opens in a new tab)), not a standard. Google says it does not use it for Search, and that you don't need special AI files to appear in its AI features (Google's guide (opens in a new tab)). Some AI tools and coding agents read it, so it's worth 8 of the 100 points here. It won't move your Google rankings.

How this check works

Our server fetches seven files from your site, the way an AI crawler would: no JavaScript, no cookies, no login. It reads one page, the one you enter, and takes up to about 20 seconds.

It says who it is. Its user agent is NazmcAIVisibilityChecker, never GPTBot or Googlebot: we don't pretend to be anyone else. To tell you how each AI crawler is treated, we read the rules your robots.txt sets for it. A site that treats crawlers differently in other ways, by firewall or by user agent, may answer them differently than it answered us.

  1. 1

    The page you enter

    As plain HTML. We never run its JavaScript.

  2. 2

    The same page, asking for markdown

    With Accept: text/markdown, to see if your server offers a markdown version.

  3. 3

    /robots.txt

    To read the rules for every AI crawler in the table.

  4. 4

    /llms.txt

    The proposed summary file for AI tools.

  5. 5

    /llms-full.txt

    Its optional long version. Not having one never costs points.

  6. 6

    The page's address plus .md

    A markdown copy of the page, if you publish one (/blog/post.md, or /index.md for the homepage).

  7. 7

    Your sitemap

    The one robots.txt names, or /sitemap.xml.

Who's crawling

Who's crawling: training, search and user agents

AI companies send different crawlers for different jobs, and robots.txt lets you answer each one on its own line. Blocking most training crawlers doesn't take you out of AI search: that is a different crawler. Google-Extended is the one exception, below.

  • Training crawlers

    Your call

    Collect pages to train future AI models.

    • GPTBot
    • ClaudeBot
    • anthropic-ai
    • Google-Extended
    • Applebot-Extended
    • CCBot
    • Bytespider
    • Meta-ExternalAgent
    • Amazonbot
    • cohere-ai

    If you block them: Keeps your pages out of future training. It doesn't remove what was collected before, and for most of these it doesn't take you out of AI search. Google-Extended and Applebot-Extended aren't separate crawlers: they're robots.txt switches for whether Google and Apple may use what their normal crawlers fetch for AI. Google-Extended is the exception: it also decides whether Gemini apps can use your pages to ground their answers. Google says blocking it doesn't affect Google Search.

  • AI search crawlers

    Keep allowed

    Build the indexes that AI search tools answer from and link to.

    • OAI-SearchBot
    • Claude-SearchBot
    • PerplexityBot
    • Amzn-SearchBot
    • DuckAssistBot
    • YouBot

    If you block them: Takes you out of that tool's search answers and citations. OpenAI says a site that opts out of OAI-SearchBot isn't shown in ChatGPT search answers, though it can still appear as a plain link. If being found in AI answers matters to you, keep these allowed.

  • Fetches for a person

    Usually allow

    Open a page because someone asked an AI assistant to, right then: "summarise this page", "compare these two".

    • ChatGPT-User
    • Claude-User
    • Perplexity-User
    • Meta-ExternalFetcher
    • Amzn-User
    • MistralAI-User

    If you block them: Stops the assistant opening your page when someone asks it to. Both OpenAI and Perplexity say robots.txt may not apply to these, because a person asked for the page.

One line per crawler

A crawler follows the group in robots.txt that names it, and only falls back to the rules for every crawler (User-agent: *) when none does. So a site can say no to training and yes to search at the same time, as in this example. Rules are matched the way the robots.txt standard, RFC 9309 (opens in a new tab), describes.

The vendors' own pages: OpenAI (opens in a new tab), Perplexity (opens in a new tab), Google (opens in a new tab).

# Keep pages out of OpenAI's training, stay in ChatGPT search
User-agent: GPTBot
Disallow: /

User-agent: OAI-SearchBot
Allow: /

User-agent: ChatGPT-User
Allow: /

Questions

Want a scan or your email deleted? Ask us and we'll remove it.

Does this check whether ChatGPT mentions my brand?

No. It checks whether AI systems can reach and read your site, which has to be true before they can quote it. To see whether you're mentioned, ask the assistants yourself: put the questions your customers ask into ChatGPT, Claude, Perplexity and Google's AI Mode, in a private window, and note who gets named and which pages are linked. Answers change from one run to the next, so try a few times and repeat it every month or so.

What is llms.txt and does Google use it?

llms.txt is a proposal (llmstxt.org), not a standard: a short markdown file at /llms.txt that lists your key pages with a line about each, for AI tools to read. Google says it does not use it for Search, and that you don't need special AI files to appear in its AI features. Some AI tools and coding agents do read it, and it takes minutes to write, so this check gives it some points, but it won't move your Google rankings.

Do AI crawlers run JavaScript?

Most don't. The crawlers from OpenAI, Anthropic and Perplexity fetch your raw HTML and don't run JavaScript, so anything your page draws in the browser after it loads is invisible to them. Google renders JavaScript, so Google Search and its AI features can see it, though sometimes later. This check reads the raw HTML, the way those crawlers do. The fix is to server-render or pre-render your main content.

Should I block GPTBot?

That's your call. GPTBot collects pages to train OpenAI's models. Blocking it keeps your pages out of future training, and it doesn't take you out of ChatGPT search, which uses a different crawler, OAI-SearchBot. If you want to be found and cited in AI answers, keep the search crawlers (OAI-SearchBot, Claude-SearchBot, PerplexityBot) allowed, whatever you decide about training.

Is my data stored?

We store the address you checked, the result and a scrambled form of your IP address, so the share link works and so we can stop abuse. We store your email only if you ask for the report by email. We don't sell any of it. To have a scan or your email deleted, ask us through the contact page.

Why is my score different from other tools?

Every tool weighs different things. This one runs 22 checks in four areas on one page, reads it as raw HTML without running JavaScript, and skips rather than fails anything it couldn't reach. A tool that renders JavaScript, crawls many pages or tracks brand mentions will give a different number. Use the fixes, not the score: they say exactly what to change.

Does a high score mean AI will cite my site?

No. A high score means nothing technical stops AI systems from reaching and reading your site. Whether they cite you depends on what your pages say, how well they answer the question being asked, and who else links to you. The score clears the way; the content does the rest.

Check your homepage first, then the page that sells

Free, no login, up to 20 seconds a check. Run it again after each fix.

Check a site