GPTBot
AI / LLMby OpenAI
GPTBot
OpenAI's crawler for collecting training data for its GPT language models. Content crawled by GPTBot may be used in future ChatGPT training datasets. Respects robots.txt. Distinct from the ChatGPT-User bot which retrieves live web browsing results.
Respects robots.txt
Yes
Can be blocked
Yes
Crawl-Delay support
No
Type
AI / LLM
Purpose
Training data collection for GPT models (ChatGPT, GPT-4)
SEO Impact
Blocking GPTBot prevents your content from being used in GPT training data. It does not affect ChatGPT's live browsing feature or AI Overviews — those use different bots.
User-Agent String
GPTBot/1.1
robots.txt Control
Add "User-agent: GPTBot" with "Disallow: /" in robots.txt. Also block "ChatGPT-User" for live browsing.
Block
User-agent: GPTBot Disallow: /
Allow (default)
User-agent: GPTBot Allow: /
Official Documentation
Test your robots.txt against GPTBot
Check which paths are blocked or allowed for each user-agent