itsbydavid
    Free AI visibility checker

    Can ChatGPT read your website?

    Paste your URL. I will tell you exactly what GPTBot, ClaudeBot and PerplexityBot receive when they ask for your page, and give you an AI visibility score in seconds. Free, no signup.

    AI visibility check

    What this AI visibility checker measures

    Server-side rendering

    Whether your content is in the raw HTML, or only appears once JavaScript runs. AI crawlers do not run it.

    Structured data

    Whether valid schema describes your business, your services and your prices.

    llms.txt checker and robots.txt check

    Whether you have a map for AI systems, and whether your robots.txt turns them away without your knowing.

    Retrievable content

    Whether a model can lift your answers and attribute them.

    Reading the result

    What your score actually means

    What does my score mean?

    It is out of 100, and the band matters more than the number. Under 20 means AI crawlers get essentially nothing from your page. Under 45 means something is badly broken. Under 70 means they get part of it. Under 85 is working. Above that, the machinery is doing its job and the remaining gains are in content, not plumbing.

    What does it mean that a crawler receives 12 words?

    It means your page is built in the browser and the server sends an almost empty shell. A person sees a full page because their browser runs the JavaScript. GPTBot, ClaudeBot and PerplexityBot do not run it, so they receive what the server sent, which is those 12 words. This is the single most common finding and usually the most fixable.

    Why is my score capped?

    Because a page can score well on everything measurable and still have almost nothing to read. Under 300 words the score is held at 45, and under 100 words with no heading at all it is held at 15. Valid schema on an empty page is still an empty page, and letting the markup carry the number would be flattering you into ignoring the real problem.

    Does a high score mean ChatGPT will cite me?

    No, and this is the distinction the whole tool is built on. A high score means the machinery works: crawlers receive your content, your markup is valid, nothing is turning them away. Whether a model names you depends on your authority and what else exists on the topic. Legibility is necessary, not sufficient.

    What should I fix first?

    Whatever is blocking delivery, always. In order: a page that renders empty to crawlers, then a robots.txt turning away crawlers you did not mean to turn away, then missing or invalid schema, then thin content. The first two are usually days of work and worth more than everything below them combined.

    Why does another tool give me a different score?

    Because most of them measure something else. This one runs deterministic HTTP checks, so two runs of the same page give the same number. Tools that ask language models whether they mention you are measuring citation, which is a different question and is not reproducible run to run. Neither is wrong; they are not comparable.

    What it does not measure

    It measures legibility, not citation

    Being readable is not the same as being cited.

    This checks whether the machinery works. Whether ChatGPT names you is a different question, and you have to ask the models.

    Why the distinction matters

    It works the other way round too. Sites that block AI crawlers still get cited, because much of the retrieval happens from search results rather than from your server.

    Any tool that blurs the two is selling you something.

    It also cannot run a browser. So it reports what a crawler receives as measured, and what a browser would see as an inference, labelled as one.

    This is an AEO checker in the narrow sense: it checks the machinery, not the citations.

    If the score is bad

    What happens then

    Then you know something your competitors probably do not know about themselves. The tool says what is broken. The call says whether it is worth fixing.

    No pitch. If your business gains nothing from this, I will say so, and I will not charge you for hearing it.

    Book a 20-min call

    Questions

    Common questions

    Is it really free?

    Yes. No signup, no card, no email. It runs deterministic HTTP checks, so it costs me almost nothing.

    Does it ask ChatGPT about me?

    No. It never calls a language model. Every check is an HTTP request and a parse, which is why it is instant and why two runs give the same score. Asking models whether they mention you is non-deterministic and expensive. That is the paid audit.

    Will this hammer my server?

    No. About twenty requests, in two batches of seven rather than all at once. Two live scans per domain per hour globally, and a repeat of the same page inside the hour is answered from cache without touching your server.

    Can I stop it scanning my site?

    Yes, one line in robots.txt. The bot documents itself, including how to block it by name.

    How the scanner identifies itself · Crawler facts last verified 9 August 2026