Free · instant · no signup

Check whether AI
can read your website

Type your website address. We look at it the way ChatGPT, Claude, Perplexity and Google do, then tell you what they can and cannot see. Every fix comes with text you can paste and a short guide you can follow yourself.

Runs the six checks that held across every industry in the research. Picking an industry adds a few more rows to the report. It never changes the score.

We read four public files: your homepage, /robots.txt, /sitemap.xml and /llms.txt, plus whatever sitemap your robots.txt points at if it lives somewhere else. Every one of them is already open to any crawler on the internet. We send the request carrying the user agent ChatGPT's crawler sends, because that is the only way to find out what it would see, then repeat it as a browser so we can tell a real block apart from a site that is down. We keep only the domain, the score and the date, and we do not ask for your email.

What this checks, and what it does not

What it checks

Whether an AI can reach your site, read the words on it, tell who you are, and find your other pages. Twenty-nine checks, run on your live site, right now, the way a real AI crawler would. Twenty-three of them are scored, with the same weights as the paid report, so the two cannot disagree. Every check has a plain-words guide in the knowledge center. Under the score, a short list of deeper checks reports what recent research found about the pages AI cites: how the page opens, how it is sectioned, whether the server answers fast, whether old addresses still work. Pick your industry and a few more rows appear. None of them move the score.

What it does not check

Whether AI actually recommends you. That is a different question with a different answer, and no technical scan can tell you. To answer it, somebody has to ask the AIs the questions your buyers ask and read what comes back. That is the Snapshot.

Why our score is not the same as everyone else's

Nothing speculative is scored

Some audits count llms.txt, MCP endpoints and Markdown content negotiation against you. Almost no site on earth has those, so almost every site lands in the fifties and gets told it is failing. We name them, say how early each one is, and give them zero points. We do test your site for llms.txt. The other three we do not test at all, and the results say so rather than implying we looked.

A crawler your robots.txt does not block is a pass

A robots.txt that does not mention OAI-SearchBot is allowing it. Showing that as a row of red warnings is a scare tactic. We show it green, because it is allowed.

Every finding says how sure we are

Proven means a vendor documents it or we watched it happen on your site. Likely means the reasoning is sound and the evidence is indirect. Emerging means nobody knows yet, so it earns no points.

After the check

Most of what the check finds, you can fix yourself. Each item links to a plain-words guide in our knowledge center. Most fixes take about ten minutes. If you would rather not, Setup puts the files and settings in place for you, for $350 in 3 business days. If the check comes back clean, the technical side is done, and the next question is whether your name comes up when your buyers ask an AI who to hire. The Snapshot is $750 and answers that one with about 25 real questions and screenshots of the actual answers, in 3 business days.