GPTBot
OpenAI · Training crawlers · User agent + IP range
About this crawler
Crawls public content that may be used to train OpenAI's generative AI foundation models.
Usually means: Public content is being collected that may end up in model training data or large web datasets.
User agent
GPTBot
MetricRaven matches this token anywhere in the request's user agent, ignoring case. Operators add version numbers and browser text around it, for example "GPTBot/1.0".
IP verification
Verified when the user agent and the IP range both match
1User agent
The request's user agent contains
GPTBot.2Request IP
Your server-side tracker sends the IP that made the request.
3Range check
The IP is compared with the ranges OpenAI publishes, refreshed twice a day.
Range sources
GPTBot IP ranges
Control it in robots.txt
User-agent: GPTBot Disallow: /private/
Well-behaved crawlers re-read robots.txt within a day or so.
Questions
- What is GPTBot?
- Crawls public content that may be used to train OpenAI's generative AI foundation models. MetricRaven groups it under Training crawlers: crawlers collecting public pages that may be used to train AI models.
- How do I verify that a request really comes from GPTBot?
- Check both the user agent and the IP. Requests from GPTBot contain "GPTBot" in the user agent and come from the IP ranges OpenAI publishes (https://openai.com/gptbot.json). A request with the name but another IP is likely spoofed.
- How do I block or allow GPTBot?
- Add a robots.txt group for "GPTBot" with the paths to allow or disallow. Fetchers that act for a user may ignore robots.txt; block those at your server or firewall instead.
- How can I see when GPTBot visits my website?
- Most crawlers never run JavaScript, so browser analytics miss them. MetricRaven's @metricraven/ai-crawl package reports crawler requests from your server, verifies their IPs against published ranges, and shows which pages each crawler reads.
See when AI assistants and crawlers read your site
Add one line to your server and MetricRaven shows which assistants, search engines and training crawlers request your pages, which ones are IP-verified, and the URLs they expected to find, including 404s.