WebmasterID logoWebmasterID
Robots & crawl control

How to block GPTBot in robots.txt

If you do not want OpenAI's training crawler fetching your site, you can disallow GPTBot in robots.txt. This page gives the exact rule, clarifies that it does not affect ChatGPT-User or OAI-SearchBot, and is honest about the limits of robots-based blocking.

Verified against primary sources

The rule

Add this group to your robots.txt to disallow GPTBot across the whole site:

User-agent: GPTBot Disallow: /

To block only part of the site, list specific paths instead of /, and use Allow to carve out exceptions.

What it does and does not affect

This rule targets only the GPTBot token. It does not affect ChatGPT-User (real-time browsing on a user's behalf) or OAI-SearchBot (OpenAI search features). If you want to restrict those too, add their tokens as separate groups. It also has no effect on non-OpenAI crawlers.

The limits

robots.txt is honoured by compliant crawlers; OpenAI states GPTBot respects it. But robots.txt cannot force compliance, and a disallow does not retroactively remove content already used. Treat it as a forward-looking request, and verify suspicious GPTBot requests by IP.

How it appears in analytics and logs

After adding a GPTBot Disallow, compliant GPTBot requests to the blocked paths should stop. Continued requests claiming to be GPTBot warrant IP verification against OpenAI's published ranges.

Diagnostic use case

Disallow GPTBot site-wide (or on specific paths) while leaving other OpenAI tokens and other crawlers under their own rules.

What WebmasterID can help detect

WebmasterID shows GPTBot crawl activity before and after your change, so you can confirm the block took effect for the compliant crawler — and flag any client ignoring it.

Common mistakes

Privacy and accuracy notes

Blocking a crawler is a publishing-policy choice, not a privacy mechanism. The rule itself is public in your robots.txt.

Related pages

Sources and verification notes

Last reviewed 2026-06-24. Facts are checked against primary/official sources where available; uncertain specifics are marked “Data not yet verified” rather than guessed.