AI crawlers / GPTBot
GPTBot: what it is and how to block or charge it
GPTBot is run by OpenAI. OpenAI's crawler for collecting public web content that may be used to train its models.
| Operator | OpenAI |
|---|---|
| Purpose | Collects content for AI training |
| User-agent contains | GPTBot |
| Official IP list | openai.com/gptbot.json (18 ranges). Verify an address |
| Same operator | OAI-SearchBot, ChatGPT-User |
Block GPTBot in robots.txt
User-agent: GPTBot
Disallow: /robots.txt is a request, not a lock: well-behaved crawlers follow it. To block by user agent at the server, match GPTBot in the User-Agent header.
Or charge GPTBot per page
Instead of blocking it, answer GPTBot with HTTP 402 Payment Required and a price. Crawlers and agents that support x402 pay a fraction of a cent in USDC straight to your wallet and get the page; people and search engines browse free. Set it up free on pay-per-crawl in about five minutes.
Is GPTBot reading your site?
See the full list of AI crawlers. Facts here follow each operator's own documentation; tell us if one changes.