What is GPTBot?

GPTBot is OpenAI's web crawler, identified by the user-agent string GPTBot, that visits publicly accessible websites to collect content for training and updating ChatGPT's knowledge base. Blocking GPTBot in your robots.txt prevents ChatGPT from indexing or citing your content.

GPTBot operates under two modes: training crawls (reading content to update model weights, governed by GPTBot in robots.txt) and retrieval crawls via OAI-SearchBot (reading content to answer live queries). The distinction matters: you may want to allow retrieval while restricting training use of proprietary content.

As of 2026, GPTBot and OAI-SearchBot together form the backbone of ChatGPT's citation pipeline. Sites that block both are invisible to ChatGPT in AI search results, regardless of traditional SEO performance. The robots.txt configuration to allow retrieval while blocking training is: User-agent: OAI-SearchBot / Allow: / followed by User-agent: GPTBot / Disallow: /.

To check whether your site allows or blocks GPTBot, use CiteFuel's free AI Crawler Access Checker.

Related

Frequently asked questions

Is GPTBot the same as ChatGPT?

No. GPTBot is the crawler that feeds ChatGPT's knowledge. ChatGPT is the AI assistant users interact with. Blocking GPTBot means ChatGPT can't cite you; allowing it is necessary but not sufficient — content must also be citable at the passage level.

Can I block GPTBot training but allow ChatGPT citations?

Yes. Block GPTBot (training) while allowing OAI-SearchBot (retrieval). CiteFuel's AI Crawler Checker generates the exact robots.txt block for this configuration.

See why AI ignores your site. Then fix it today.

Free 23-check audit. No card. No login. Just a URL — results in ~90 seconds.

Audit my site free →