Check what AI search can actually read on your website
We fetch your page the same way ChatGPT, Perplexity and Google's AI crawlers do, without running any JavaScript, and show you what comes back. It takes about ten seconds and the result appears right here.
What the check looks at
The same three tests an AI engine applies before it names a business in an answer.
Can AI read you
Whether your robots.txt lets GPTBot, ClaudeBot, PerplexityBot and the rest in by name, and whether your words exist in the HTML or only appear after JavaScript runs. This is where most sites quietly fail, and the owner never finds out.
Do you answer the question
Whether the page has a real title, one clear heading, enough written content to quote, and headings shaped like the questions customers actually ask. Engines lift answers. A page with no answers on it gives them nothing to lift.
Can AI quote you
Whether your business is described in structured data, whether your phone number is machine-readable, and whether your sitemap and llms.txt exist. This is how an engine states who you are without having to guess from prose.
What a machine cannot tell you
needs a personThis check reads your page. It cannot read the rest of the internet, and it will not pretend to. Four things decide whether an engine actually names you, and none of them can be measured from a single fetch. Mike goes through these by hand when you send your site over.
Questions about the check
Fetch your page the way an AI crawler does, without running JavaScript, and see whether any words come back. Two things decide the answer. First, your robots.txt has to allow the AI crawlers by name, because GPTBot, ClaudeBot and PerplexityBot each check for their own rules. Second, your page text has to exist in the raw HTML rather than being drawn in by a script after the page loads. If either one fails, the engine sees a blank page no matter how good the writing is.
Most AI crawlers fetch the raw HTML and do not run JavaScript the way a browser does. A site built so that its text is drawn in by a script after loading looks complete to you and close to empty to them. The fix is to send the words in the HTML itself, through server rendering or a static build, so the text is there the moment the page is fetched.
It is a plain text file at the root of your site that tells AI crawlers what your business does, where it operates and which pages matter. It works like robots.txt, but instead of setting rules it gives a short summary an engine can read in one request. It is not required and no engine promises to use it, which is why this check reports it as a small point rather than a critical one.
No. The score covers what can be measured from your page: whether crawlers are allowed in, whether your text is readable, and whether you state plainly who you are and what you answer. Being named in an actual answer also depends on your reviews, your listings and how many other sources agree about you. None of that can be read off a single page, so this check does not pretend to score it.
No. The result appears on this page as soon as the check finishes, and every finding is visible without signing up for anything. Your details are only asked for at the end, and only if you want Mike to go through the parts a machine cannot check, like your listings and reviews.
The result is cached for fifteen minutes so that checking the same site twice does not fetch it twice. Nothing is written to a database and no report is kept against your name. If you choose to send the findings to Mike at the end, the address and the report travel in that message and nowhere else.
What this looks like in your trade
The questions customers ask AI, and the listings it cross-checks, differ by industry.
Want the part a machine cannot check?
Run the check above first, then send it over. Mike goes through your listings, your reviews and what the engines say about you today, and replies with what he finds. No charge and no sales sequence.