Check whether AI crawlers can reach your homepage.
Three named checks on your homepage: whether robots.txt lets in the crawlers that cite sources, whether the page asks not to be quoted, and how much text arrives before JavaScript runs. A failed check comes with a plain fix and a prompt you can paste into a coding tool. It takes about a minute.
No sign-up, no email. The address is not saved, and no copy of the report reaches me. What happens to it.
Want speed, technical SEO and visitor paths as well? Run the full site check.
Looking at the site.
Loading the homepage the way a browser would and reading robots.txt. This runs the same audit as the full site check, so it also asks PageSpeed for speed numbers. That part is the slow one: it runs a real browser on Google's hardware.
Three checks, counted by name.
You get how many passed, not a score. llms.txt and structured data are shown as notes.
Crawlers that cite sources
Whether robots.txt lets in the crawlers that fetch pages to cite them in AI answers, asked for the homepage address. Rules with a * in the path are not understood, and a firewall or CDN block on crawlers does not show. Other AI crawlers are shown apart and do not count.
The homepage robots tag and header
Whether the first robots meta tag or the X-Robots-Tag header says noindex, none, nosnippet or max-snippet:0 for all crawlers, a search crawler or a crawler that cites sources. A rule in there for another AI crawler or a name I do not know is shown as a note. A separate meta tag for one crawler by name, or a second robots tag, is not read. If a run does not return both the tag and the header, this reads not measured unless one of them already shows a block.
Readable text in the HTML
How many words are in the homepage HTML before JavaScript runs. The check asks for 150 or more. If a run returns no word count, this reads not measured.
What it will not do. This shows what robots.txt and the homepage HTML say, not whether an AI answer will cite the site.
Questions about the AI visibility check.
What does it read?
Your robots.txt, your llms.txt if there is one, your sitemap, and the HTML and headers the server sends for your homepage before JavaScript runs. It also reads a few pages your homepage links to, such as pricing and contact pages. Three things can pass or fail: the crawlers that cite sources, the homepage robots tag and header, and the words in the homepage HTML. llms.txt and structured data are notes rather than checks, because Google says its generative AI search features need neither, though structured data still earns rich results. Anything the check could not read on a run is marked not measured, and it never counts as a pass. The crawler check asks robots.txt about the homepage address only, so a block on other sections does not show. Its parser does not understand a * in a path, such as Disallow: /*, or a byte order mark at the top of the file, so a block written that way can pass. It fetches pages the way a browser would, so a firewall or CDN that blocks crawlers by name does not show either. It reads the first robots meta tag and the X-Robots-Tag header. A rule inside either that names a crawler, such as googlebot: noindex, fails for a search crawler or a crawler that cites sources, and is a note for another AI crawler or a name I do not know. A separate meta tag for one crawler by name, such as googlebot, is not read, so a block written that way does not show.
Why is there no score?
A score out of 100 would claim a precision this does not have. Nobody publishes what earns a citation, so any weighting would be one I made up. A count of named checks says exactly what passed and what to fix. Checks that could not be measured are left out of the count and named.
Does it predict whether AI answers will cite my site?
No. It shows what robots.txt and the homepage HTML say. Passing every check does not earn a citation, and failing one does not rule it out. I cannot see what any AI product will choose to cite.
What is the difference between crawlers that cite and other AI crawlers?
Crawlers that cite, such as OAI-SearchBot, ChatGPT-User, PerplexityBot and Claude-SearchBot, fetch pages so an answer can cite them. Other AI crawlers feed models and products: GPTBot, ClaudeBot and CCBot collect text for training, and Google-Extended covers training and grounding for Gemini. Blocking the first group fails the check. Blocking the second is a choice about your content rather than a fault, so it is shown and never counted.
What happens to the address I type in?
If you type a deeper address, only the site part of it is sent, because this check reads the homepage. It goes to the machine that runs the check, which fetches your site the way a browser would. It also goes to Google PageSpeed Insights, because this page runs the same audit as the full site check and that audit asks Google for speed numbers. There is no email and no sign-up. It is not written to a database, the check keeps no log of it, and no copy of the report comes to me. The privacy page sets out exactly what happens to it.
Let me plan your next step.
I can scope a migration. Send a brief and I will reply by email with next steps.