What is GPTBot?
GPTBot is a ai or training crawler operated by OpenAI. Use this crawler profile to identify user-agent tokens, operator signals, platform hints, and recommended handling.
GPTBot
OpenAI · AI or training crawler
- Version
- 1.4
- First seen
- May 21, 2026
- Last seen
- May 24, 2026
- Confidence
- Known user-agent token
What is GPTBot?
GPTBot is a ai or training crawler operated by OpenAI. GPTBot is likely used by OpenAI to collect or evaluate public web content for AI, language, or training systems.
GPTBot matched a known crawler token in the user-agent string, but this page alone does not prove IP ownership.
How to identify GPTBot in logs
Search server logs for GPTBot. Matching those tokens is useful for discovery, but IP verification is still recommended before trusting the identity.
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; GPTBot/1.4; +https://openai.com/gptbot)
Use this as a web-crawler lookup reference for identifying how this user-agent presents itself in server logs.
User-agent signals
- Product tokens
- Mozilla/5.0, AppleWebKit/537.36, Gecko, GPTBot/1.4
- Documentation
- https://openai.com/gptbot
- Contact
- None found
- Platform
- Unknown · Unknown
- Browser profile
- Unknown · WebKit
- Browser-like UA
- Yes
- HTTP library
- Unknown
- Spoof risk
- Medium
Questions answered by this crawler profile
What is GPTBot?
GPTBot is likely used by OpenAI to collect or evaluate public web content for AI, language, or training systems.
Who operates GPTBot?
GPTBot is associated with OpenAI.
How do I identify GPTBot in logs?
Search server logs for GPTBot. Matching those tokens is useful for discovery, but IP verification is still recommended before trusting the identity.
Should I allow GPTBot?
Verify IP ownership or behavior before making security decisions because user-agent strings can be spoofed. Review your robots.txt and AI crawler policies, then allow, block, or rate-limit it based on your content usage preferences.
Does GPTBot respect robots.txt?
Robots.txt compliance cannot be proven from a user-agent string alone. Check the crawler operator documentation and your own logs before assuming behavior.