Which bots do you check?

The major AI assistant and search crawlers, from a maintained inventory of known AI user-agents. The check reads what your robots.txt declares for each one at your site root.

Why does a bot show up as unknown?

Because we could not verify your robots.txt: the request was challenged or blocked, timed out, or returned something that is not a rules file. Rather than guess "allowed," we show unknown with the reason. A firewall or bot-management layer can also override robots.txt entirely, which is exactly why an unconfirmed result is never dressed up as an all-clear.

Does blocking crawlers hurt my visibility?

It can. If the assistants cannot read you, they answer questions about your brand from other sources. But blocking is sometimes deliberate, especially for training crawlers, so the goal is to see your real state and decide on purpose, not to open every door by default.

Free tool · No signup

See which AI crawlers can actually reach your site.

AI assistants only recommend brands they can read. This checks whether the major AI crawlers can actually reach your site, and tells you the honest truth when it cannot tell.

At a glance

Who it's for

Anyone who owns a website and wants AI assistants to find and recommend it. No signup, no crawler expertise needed.

What it checks

Whether each major AI crawler is allowed, blocked, or unknown at your site, read straight from your robots.txt the way that bot reads it.

Why it matters

AI assistants answer from what they can reach. Seeing your real access state is the first thing to fix before anything else.

What the output means

Three answers, and one of them is "we could not tell."

For each AI crawler the check reads your robots.txt the way that bot does and reports one of three states: allowed, blocked, or unknown. Allowed and blocked are what the site declared. Unknown is the honest one.

When your robots.txt cannot be verified, because it was challenged, failed, or came back as something that is not a rules file, most checkers quietly call that "allowed." We never do. A verdict we could not confirm is shown as unknown, with the reason, because a false all-clear is worse than an honest "we could not tell."

A blocked verdict is a choice, not always a bug. Plenty of brands block training crawlers on purpose. The point is to see the real state on purpose, rather than assume the doors are open.

This is the free checker. The full product is Jinn Agents.

What you ran above is one public check, free and no account needed: a small taste of the engine. Jinn Agents is the full product, the brand-grounded engine behind it. Once your brand is Connected there, this same read runs against your whole surface on every review, and the fixes ship as remediation playbooks with the exact robots.txt and markup to change, not a PDF of problems. It runs on demand, when a review runs, against the pages that matter.

See Jinn Agents

Keep going.

Free toolllms.txt GeneratorWrite a correct llms.txt from your live site in about ten seconds, with real names slotted in and TODO markers where only you know the truth.Free toolAI Visibility SnapshotReachable is step one. Run the free audit to see how AI engines actually answer questions about your brand today.

DictionaryAI crawlerThe bots that read the web for AI assistants. If they cannot read you, the assistants answer questions about your brand from everyone else.

DictionaryGEOGenerative engine optimization: the practice of making sure AI answer engines can find, read, and correctly represent your brand.

This tool also ships as an installable skill, so your own agent can run the same check locally:

npx skills add jinn/agent-access-checker

MIT-licensed. The skill carries the method; nothing in it calls Jinn.

Questions

What people ask.

Which bots do you check?: The major AI assistant and search crawlers, from a maintained inventory of known AI user-agents. The check reads what your robots.txt declares for each one at your site root.
Why does a bot show up as unknown?: Because we could not verify your robots.txt: the request was challenged or blocked, timed out, or returned something that is not a rules file. Rather than guess "allowed," we show unknown with the reason. A firewall or bot-management layer can also override robots.txt entirely, which is exactly why an unconfirmed result is never dressed up as an all-clear.
Does blocking crawlers hurt my visibility?: It can. If the assistants cannot read you, they answer questions about your brand from other sources. But blocking is sometimes deliberate, especially for training crawlers, so the goal is to see your real state and decide on purpose, not to open every door by default.

See your real access state. Then fix it on every page.

Run the check above to see where the AI crawlers actually stand today. When your brand is Connected, this runs across your whole surface and the fixes ship as remediation playbooks.

See how brands connect