ChatGPT-User: when ChatGPT visits a page for a user
ChatGPT-User fetches pages when a ChatGPT user or custom GPT asks for them. Why robots.txt may not apply, its IP list and how to verify it.
Updated
ChatGPT-User is the user agent ChatGPT uses when it visits a web page because a user asked for something that needs it, in ChatGPT or a custom GPT. OpenAI says it is not used to crawl the web automatically, and that because these visits are started by a user, robots.txt rules may not apply.
| Owner | OpenAI |
|---|---|
| Purpose | User-triggered |
| robots.txt token | ChatGPT-User |
| Follows robots.txt | May not apply (user-initiated) |
| Published IP ranges | openai.com/chatgpt-user.json |
| If you block it | ChatGPT may fail to open your pages when a user asks it to. Not used for search inclusion. |
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai.com/botWhat does ChatGPT-User do?
It fetches a specific page at a user’s request: reading a URL someone pasted, opening a page during a task, or calling an action in a custom GPT. One request generally means one person wanted something from your page. It is different from OAI-SearchBot, which crawls to build ChatGPT’s search index, and GPTBot, which collects training data.
Does it follow robots.txt?
OpenAI’s wording is that robots.txt rules “may not apply” to user-initiated actions. So a Disallow for ChatGPT-User is not a reliable block. If you must stop these visits, block the IP ranges OpenAI publishes for ChatGPT-User at your firewall or CDN.
User-agent: ChatGPT-User
Disallow: /private/What happens if I block it?
ChatGPT can’t open your page when someone asks it to, so it answers without your content or tells the user it couldn’t access the site. That is usually a lost chance to be described accurately, by the people most interested in you. It doesn’t change whether you appear in ChatGPT search.
How do I tell the real ChatGPT-User from a fake?
Check the IP address against openai.com/chatgpt-user.json. Requests with this user agent from other addresses are not from OpenAI.
Should you block ChatGPT-User?
Rarely. Block it only for areas you don’t want any automated agent to read, and protect truly private pages with a login, which no fetcher can pass. For public pages, letting users’ ChatGPT sessions read them helps those users get accurate answers about you. See your full AI bot picture with the AI crawler checker.
Sources
Common questions
Is ChatGPT-User a crawler?
OpenAI says it is not used for crawling the web in an automatic fashion. It visits pages because a person using ChatGPT or a custom GPT took an action that needs them.
Does ChatGPT-User decide whether I appear in ChatGPT search?
No. OpenAI says ChatGPT-User is not used to determine whether content may appear in search. Use OAI-SearchBot in robots.txt to manage search.
Why do I see ChatGPT-User visits in my logs?
Someone asked ChatGPT something that led it to open your page, for example by pasting your URL or asking about your product. Those visits are a sign people are researching you through ChatGPT.