# ChatGPT-User

## Quick answer

ChatGPT-User is the OpenAI agent that fetches a specific web page when a person asks ChatGPT, or a Custom GPT, a question that requires visiting that URL. It isn't a general web crawler; it only acts on a live user's request, one page at a time.

## What it does

ChatGPT-User fires when someone using ChatGPT, or a Custom GPT built on top of it, asks a question that requires fetching a specific page rather than answering from what the model already knows. OpenAI is explicit that this isn't used for crawling the web automatically; there's no background sweep of a site the way a search index crawler works, only one-off fetches tied to a real person's request in the moment. Because the fetch is triggered by that person's own action, OpenAI notes that robots.txt rules may not apply in the usual way, since the request behaves more like a user's browser asking for a page than an automated crawl. OpenAI publishes ChatGPT-User's IP ranges separately so site owners can confirm genuine traffic.

## Facts

- Operator: OpenAI
- Robots token: ChatGPT-User
- User agent string(s): Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai.com/bot
- Purpose: user-triggered-fetch
- Respects robots.txt: not stated
- IP ranges: https://openai.com/chatgpt-user.json

## Should you block it?

Blocking ChatGPT-User can stop ChatGPT from fetching a page when a user directly asks it to look at that URL, which matters for a brand that wants its pages readable when someone pastes a link into ChatGPT or a Custom GPT references it. Because OpenAI says robots.txt rules may not fully apply here, since the fetch is user-initiated rather than automated, a disallow rule is a signal of intent more than a guaranteed block. This bot is separate from GPTBot's training crawls, so the decision isn't about AI training data; it's about whether a brand's own pages are reachable at the moment a real person is asking ChatGPT about them. MarketHQ tracks how brands appear in ChatGPT answers, and a site that blocks user-triggered fetches risks ChatGPT having less current, first-party information to work with when someone asks about that brand directly.

## How to block it

Add this to robots.txt to discourage ChatGPT-User fetches:

User-agent: ChatGPT-User
Disallow: /

Because OpenAI notes robots.txt rules may not apply to this user-triggered fetcher, this is a signal rather than a guaranteed block.

## FAQ

### Is ChatGPT-User the same as GPTBot?

No. GPTBot crawls for training data automatically. ChatGPT-User only fetches a page when a real person asks ChatGPT, or a Custom GPT, about that specific URL.

### Does robots.txt reliably block ChatGPT-User?

Not necessarily. OpenAI says that because these fetches are user-initiated, robots.txt rules may not apply, unlike GPTBot's automated training crawl.

### Does ChatGPT-User power live ChatGPT search answers in general?

It supports fetching a specific page a user or Custom GPT asks about. OpenAI documents separate bots for different ChatGPT functions, each with its own rules.

## Sources

- ChatGPT-User is not used for automatic crawling of the web. Source: https://developers.openai.com/api/docs/bots
- Because ChatGPT-User fetches are user-initiated, robots.txt rules may not apply. Source: https://developers.openai.com/api/docs/bots