GPTBot

GPTBot collects public web content that may be used to improve OpenAI's generative AI models. Aiola classifies it as training because it performs automated collection rather than a user-triggered fetch.

In Aiola's taxonomy, Training means scraped for model training or dataset development.

Look for the match token inside the complete HTTP User-Agent header. Tokens can be spoofed, so use the network checks below for authentication.

Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.4; +https://openai.com/gptbot

GPTBot

OpenAI publishes an official IP range list. This cached snapshot contains 18 CIDR prefixes, last fetched 2026-10-23. Open the live source

132.196.86.0/24
172.182.202.0/25
172.182.207.0/25
20.125.66.80/28
20.171.206.0/24
20.171.207.0/24
4.227.36.0/25
52.230.152.0/24
74.7.175.128/25
74.7.227.0/25
74.7.227.128/25
74.7.228.0/25
74.7.230.0/25
74.7.241.0/25
74.7.241.128/25
74.7.242.0/25
74.7.243.128/25
74.7.244.0/25

For GPTBot, compare the source IP from your trusted edge or server against OpenAI’s current published prefixes.

python3 -c 'import ipaddress; print(ipaddress.ip_address("REQUEST_IP") in ipaddress.ip_network("132.196.86.0/24"))'

Test every current prefix in the vendor feed and re-fetch it regularly. This cached snapshot is evidence of publication, not a permanent firewall list.

OpenAI documents crawler controls for this bot family. Robots.txt can direct cooperative crawling, but it cannot prove that a request is authentic.

User-agent: GPTBot
Allow: /

User-agent: GPTBot
Disallow: /

Read official crawler documentation