AI crawlers
ChatGPT-User: the agent behind user requests.
This agent visits a page when someone asks ChatGPT about it. Here is who runs it, how to spot it in your logs, and what robots.txt does.
Short answer
What is OpenAI's user-triggered agent?
ChatGPT-User is OpenAI's user agent for visits triggered by a person. When someone asks ChatGPT or a Custom GPT a question, it may open your page. OpenAI says it does not crawl automatically and does not decide Search inclusion.
In brief
Four things to know first.
OpenAI runs it
OpenAI uses user agents for actions in its products, either automatically or when a user triggers them. This one is the user-triggered kind.
A person starts each visit
OpenAI says this agent is not used to crawl the web automatically. A user's question prompts the visit.
robots.txt may not apply
OpenAI says that because users initiate these actions, robots.txt rules may not apply to them.
Search inclusion is separate
OpenAI says to use OAI-SearchBot in robots.txt for Search opt-outs. This agent does not decide that.
What it does
What ChatGPT-User does with your page
OpenAI's bot documentation says that when users ask ChatGPT or a Custom GPT a question, it may visit a web page using this user agent.
That makes it a request-time visit, not a crawl. OpenAI lists it for certain user actions in ChatGPT and Custom GPTs.
| The user-triggered agent | OAI-SearchBot | GPTBot | |
|---|---|---|---|
| Purpose | Certain user actions in ChatGPT and Custom GPTs | Surface sites in ChatGPT search results | Crawl content that may train OpenAI's models |
| Trigger | A user asks a question | OpenAI's own crawling | OpenAI's own crawling |
| robots.txt | Rules may not apply | Controls Search inclusion | Controls training use |
Recognise it
Spot the ChatGPT user agent in logs
OpenAI publishes the full string for this agent. Match it in your logs, then compare the request IP against OpenAI's published list.
| Item | Value |
|---|---|
| User agent | Mozilla/5.0 AppleWebKit/537.36 string with OpenAI's agent token, version 1.0, and info URL +https://openai.com/bot |
| Info page | https://openai.com/bot |
| Published IPs | OpenAI's published IP list for this agent (a JSON file on openai.com) |
See how ChatGPT treats your site
Our free AI audit shows how your brand appears in ChatGPT, Claude, Perplexity and Gemini answers.
Control it
Handle ChatGPT-User in four steps
OpenAI's page does not give a robots.txt block that is certain to stop this agent. Treat robots.txt as unreliable here and decide with that in mind.
Find it in your logs
Search your logs for this agent's user agent token, then check the requesting IP against the address list OpenAI publishes for it.Do not rely on robots.txt
OpenAI says these actions are user-initiated, so robots.txt rules may not apply.Manage Search with OAI-SearchBot
OpenAI says to use OAI-SearchBot in robots.txt for Search opt-outs and automatic crawl.Manage training with GPTBot
OpenAI says disallowing GPTBot signals content should not be used to train its foundation models.
Should you block it?
What blocking it costs you
A visit happens because a user asked about your page or topic. Turning that agent away can mean ChatGPT cannot read your page for that user.
It has no bearing on Search inclusion. Per OpenAI, sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers. Keep your own rules for those crawlers separate.
Allow if you want to be readable
Users who ask ChatGPT about your page can have it opened for them.
Block only for a clear reason
Private or paid pages are one case. Check what your logs show first.
Check, then decide
Use the IP list, since a user agent string alone can be faked.
FAQ
ChatGPT-User: common questions.
Is the user-triggered agent the same as GPTBot?
GPTBot is a separate agent from the user-triggered one. OpenAI says GPTBot crawls content that may be used to train its generative AI foundation models. The user-triggered agent may visit a page when someone asks ChatGPT or a Custom GPT a question.
Does this agent decide if my site appears in ChatGPT search?
ChatGPT search inclusion does not depend on the user-triggered agent. OpenAI says that agent is not used to decide whether content may appear in Search. You manage Search opt-outs through OAI-SearchBot in robots.txt. Sites that block OAI-SearchBot are not shown in ChatGPT search answers.
Does robots.txt stop OpenAI's user-triggered visits?
Robots.txt may not stop this agent. OpenAI says that because users initiate these actions, robots.txt rules may not apply. Use OAI-SearchBot and GPTBot rules for Search and training, and expect user-triggered visits to follow different rules.
What is the user agent string for OpenAI's user-triggered visits?
OpenAI's bots page lists the full user agent string for this agent. The same page points to a JSON file of its published IP addresses, so you can check both.
How do I verify this OpenAI user-triggered request is real?
Verify a request from this agent by checking its IP against the JSON list OpenAI publishes on its bots page. A matching user agent string alone is not proof, since anyone can send that header.
See how ChatGPT reads your site.
We run your buyers' questions through ChatGPT, Claude, Perplexity and Gemini and show you the gap.