Can ChatGPT, Claude and Perplexity actually reach your site?
An open robots.txt is not proof of access. Security settings and CDN rules can refuse AI crawlers above that file, and nothing in your SEO report will mention it. This checks the response your site gives each one, right now.
Check whether AI crawlers can reach your site
We ask your homepage for its response code once per crawler, from our server. Nothing is stored.
What the codes mean
200 means the door is open and the crawler received your page. 403 or 401 means it was refused, and for the AI search group that is decisive: OpenAI's own documentation states that sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers. A 429 means you are rate-limiting that crawler, which usually ends the same way.
A 5xx or a timeout is not a verdict. It means your server did not answer this particular request, and the honest reading is to try again rather than to conclude anything.
The three groups exist because blocking AI is not one decision. Refusing model training while allowing AI search is a coherent, common choice. Refusing AI search is also a choice — it just needs to be one you made on purpose.
If a crawler was refused
- 1
Check robots.txt first
It is the one place you control directly. A disallow for the crawler's name, or a blanket disallow, is the simplest explanation and the fastest fix.
- 2
Then look at your CDN or security tooling
Cloudflare and similar services carry their own AI-bot switches that act above robots.txt. This is the usual culprit: robots.txt is wide open and the request never gets that far.
- 3
Check your service pages too
We ask for the homepage. Rules often differ deeper in the site, so repeat the check on the pages you actually want cited.
- 4
Confirm with your server logs
A name in a User-Agent proves nothing on its own. Match the requests against the IP ranges the vendors publish — that is the only version of this check that cannot be faked.
Questions about this check
- How do I know if my site is blocked from ChatGPT?
Request your own page while identifying as OpenAI's search crawler and look at the response code. A 403 or 401 means the request was refused, and OpenAI's documentation states that sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers. A 200 means the door was open for that request. This tool does exactly that for six crawlers at once; you can also do it yourself with a single curl command.
- Why does robots.txt say allow, but the crawler still gets a 403?
Because robots.txt is a request, not a gate. Anything in front of your server — a CDN, a firewall, a bot-management rule — can refuse a crawler before your site is ever consulted, and those settings often have their own AI-bot toggles that are on by default. This is the most common cause of an invisible block: the file is open and the request never arrives.
- Is a 200 proof that ChatGPT can see my site?
No. Anyone can put a crawler's name in a User-Agent header, so a 200 only proves your site does not refuse requests that claim that name. Serious protection checks the IP address instead. To confirm properly, look in your server logs for requests from the IP ranges the vendors publish. A 200 here is a good sign and a starting point, not evidence.
- Should I block AI crawlers?
It depends which ones. The crawlers that feed AI search decide whether your business can appear in answers at all, so blocking those costs you visibility. Training crawlers are a separate decision: refusing to have your content used for model training changes nothing about whether you get cited, and plenty of businesses do exactly that. The mistake is blocking all of them with one switch and not knowing which one you needed.
- Do you store the address I check?
No. The check runs when you press the button, the result is shown to you and nothing about the address is written down or sent anywhere. We count how many crawlers were refused, without the address, so we know whether the tool is useful.
Got a 403 and no idea where it comes from?
Send us the result. In 30 minutes we will find which layer is refusing the crawler — robots.txt, your CDN, or a security rule — and tell you what to change. Free, no strings.
See how we work on AI search visibility