Run a free check
AI visibility

Does ChatGPT recommend your brand?

There are two separate answers and most tools blur them. First, can an AI engine reach and read your site at all — that is robots rules, firewall behaviour, server-rendered HTML and structured data, and it is measurable from outside in about a minute. Second, does the model actually name you when asked. The free check on this page answers the first question in full, because the second is impossible without the first.

Check your domain free  How the check works

No signup for the score. 46 checks · 24 AI crawlers.

The honest version

Can anyone actually tell you whether ChatGPT recommends you?

Partly, and the part that is not measurable is worth saying out loud.

  • Answers vary between two runs of the same prompt.
  • They vary by country, by account, and by the day.
  • The API answer is not the answer in the app.
  • So a single "you rank #3 in ChatGPT" number is a guess wearing a suit.
What is solid. Whether an engine can fetch your pages, whether it gets HTML or an empty shell, whether your robots rules shut it out, and whether your identity is stated clearly enough to be resolved to you rather than to a company with a similar name. Those are facts about your site, they do not move between runs, and every one of them is a precondition for being recommended.
Why brands lose

Why would a model skip a company that clearly fits?

In our own audit of 46 checks across seventeen properties, the same four causes came up again and again.

  • The site is blocked and nobody knows. A CDN master switch turns on and locks the per-bot settings behind it.
  • Nothing renders without JavaScript. AI crawlers do not run yours.
  • The brand name is ambiguous. The model answers about a similarly named company in the same country.
  • Bing has never indexed the site. ChatGPT Search and Copilot read Bing.

See which of these applies to you

What you get

What does the free check actually return?

The whole audit, not a teaser.

  • A score out of 100 with the five weighted bands behind it.
  • 24 AI crawlers checked against your robots rules.
  • 14 of them requested with their real user-agent, so you see what your edge actually did — not what your config says it should do.
  • Every failure with the evidence we measured and the fix, written for a developer.
  • A re-check any time, free, and a comparison against your last one.
Two rules the tool holds itself to. A request we never sent is reported as not tested, never as blocked. And a server rate-limiting us is never called a block — those are different facts, and conflating them is how a checker manufactures a scary number.
After the score

What if you would rather someone just fixed it?

The audit is free and stays free. The work is the paid part, and the two are sold separately on purpose — the tool does not need to be bought to be useful, and the fixes do not need a subscription to be done.

See what the work costs  Ask a question