Can AI assistants actually read your site?
Enter your address. We read the page the way a crawler does and tell you what is missing. No signup, nothing stored.
Reads one page — your homepage unless you enter a deeper address. Takes about ten seconds.
Results
What this cannot see
What this does, and what it doesn't
It fetches one page and reads the HTML the way a crawler would. Fifteen checks: whether the content is there without JavaScript, the title, description, a single H1, canonical, Open Graph, structured data, an llms.txt, whether robots.txt is turning crawlers away, plus mobile, alt text, page weight and script count.
Because it is the most common reason a business is invisible in AI answers. Google runs JavaScript reasonably well; AI crawlers mostly do not, and several barely run it at all. A site assembled in the browser can look finished to you and close to empty to ChatGPT or Claude.
No, and nothing that reads a single page can. Rankings need live search data from outside your site. This tells you whether the page is set up to be found at all, which is the part you actually control.
Not to run the check — no signup, no email, nothing kept. The address is used to fetch the page and is not stored, so you can run it as often as you like on any site you like, including your competitors. There is one optional box at the bottom if you want the results emailed to you; skip it and nothing is asked of you at all.
It depends entirely on which checks failed. A missing Open Graph tag means a link to you looks plain when it is shared — worth fixing, not urgent. Content that is invisible without JavaScript, or a robots.txt blocking crawlers, can mean you are effectively absent from search and AI answers entirely. The report sorts the findings so the expensive problems are at the top; if the top of your list is all green, the score matters less than it looks.
Because one page is what can be checked honestly in ten seconds without a crawler, a queue and a paid data bill — and because most of what it finds is site-wide anyway. Titles and descriptions are per page, but the things that really hurt (content needing JavaScript, blocked crawlers, missing structured data) usually come from how the whole site is built. Run it on your home page, then on one service page, and you will see the pattern.
Fair question, and the reason every check tells you what it looked at rather than just a verdict. Nothing here is scored on taste — each finding is something present or absent in your HTML that you can go and confirm yourself. We have also fixed two false positives in this tool that were making sites look worse than they were, because a checker that invents problems is worth less than no checker at all.
Take the top two or three findings to whoever maintains your site — the report names the fix for each one, so it can be forwarded as-is. Most of what it flags is a small job for someone with access. If nobody maintains it, or the answer comes back as a rebuild, that is worth a second opinion before you spend anything.
Want the list fixed rather than just found?
Send the result over and you'll get a straight answer about what's worth doing, from the person who'd do it.
Let's talkNeed help fixing this?
Most of what this finds is a small job for whoever has access to your site. If nobody does, or the answer came back as a rebuild, we'll tell you honestly which findings are worth paying to fix and which can wait.
Send us your resultsNo charge for the conversation, and we'll tell you if it isn't worth doing.