AI crawler · OpenAI · Model training
GPTBot: what it is, its user agent, and how to block it
GPTBot is OpenAI's AI training crawler. Collects content that may be used to train OpenAI's foundation models. In our 13 October 2026 study, 337 of 2,332 live top sites with readable robots.txt rules (14.5%) disallow GPTBot. Part of the AI crawler list.
What GPTBot does
Operator: OpenAI. Job: Model training. Obeys robots.txt.
“Disallowing GPTBot indicates a site’s content should not be used in training generative AI foundation models.”
GPTBot user agent
robots.txt token: GPTBot (matched case-insensitively).
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.4; +https://openai.com/gptbotOpenAI calls this an example: the version number may change.
robots.txt: allow or block GPTBot
Allow
User-agent: GPTBot
Allow: /Block
User-agent: GPTBot
Disallow: /Training only, not citations. Blocking it is a training opt-out. It doesn't decide whether OpenAI's AI answers can cite you; OAI-SearchBot does that, and it has its own robots.txt group.
Once a group names GPTBot, it ignores your User-agent: * rules, so copy any paths you still want blocked into it.
How many top sites block GPTBot
Disallow GPTBot in robots.txt
14.5%
337 of 2,332 sites
In our 13 October 2026 study, 337 of 2,332 live top sites with readable robots.txt rules (14.5%) disallow GPTBot.
Sample: Tranco top 5,000, tested once on 13 October 2026 from one Indian fixed-line connection. The robots.txt share counts live sites (home page 2xx) whose rules we could read, as on the AI visibility hub. Method and CSV in the study.
Check your own site
The AI crawler checker reads your robots.txt the way GPTBot does and tests what your firewall returns. To write the rules, use the AI robots.txt generator.
Check your page: free, about 20 seconds, no sign-in
What it checks
- robots.txt for GPTBot
- Finds the group GPTBot obeys under RFC 9309 and the rule that decides your page.
- What your CDN lets through
- Requests the page as OAI-SearchBot, PerplexityBot and Claude-SearchBot and compares the answers with a browser request.
- Content without JavaScript
- AI crawlers don't run JavaScript. The audit checks how much of the page is in the raw HTML.
Free to run. Pay only for the fixes you want.
Free
Free account
Your score without signing in; with a free account, every check's result and the full fix for the top 1 issues.
★ Starter
$9 / ₹499 once
Every issue with its exact fix for one site, PDF/Markdown/JSON exports, re-audit for 30 days. Nothing renews.
Pro
$29 / ₹1,499 / month
Full reports on every site, weekly alerts when AI access breaks, competitor compare.
Agencies: white-label widget, branded reports and lead capture on Agency.
Questions
What is GPTBot?+
GPTBot is OpenAI's AI training crawler. Collects content that may be used to train OpenAI's foundation models.
What is GPTBot's user agent?+
OpenAI documents it as: Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.4; +https://openai.com/gptbot. Match on the token GPTBot, since version numbers change.
How do I block GPTBot?+
Add "User-agent: GPTBot" followed by "Disallow: /" to your robots.txt. Obeys robots.txt, per OpenAI.
Does blocking GPTBot affect AI search citations?+
Blocking it is a training opt-out. It doesn't decide whether OpenAI's AI answers can cite you; OAI-SearchBot does that, and it has its own robots.txt group.
How many websites block GPTBot?+
In our 13 October 2026 study, 337 of 2,332 live top sites with readable robots.txt rules (14.5%) disallow GPTBot.