What is ChatGPT-User?

ChatGPT-User is a bot in the AI assistants and agents category, operated by OpenAI.

Bot name
ChatGPT-User
Operator
OpenAI
Type
Bot
Respects robots.txt
Not confirmed
robots.txt user agent
ChatGPT-User
Signs requests (Web Bot Auth)
No

About ChatGPT-User

ChatGPT-User is for user actions in ChatGPT and Custom GPTs. When users ask ChatGPT or a CustomGPT a question, it may visit a web page to help answer and include a link to the source in its response.

AI assistants fetch pages on demand when a user asks a chatbot or an agent to read, summarise or act on a URL. Each request is tied to a person, but the visit carries no ad impressions or analytics. Decide per agent whether the value is worth the load.

ChatGPT-User user agent

ChatGPT-User identifies itself with this user agent string:

  • Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai.com/bot

Anyone can copy a user agent string. Transfon Security also checks where each request comes from, so a request that claims to be a known crawler but comes from the wrong network is treated as a fake, not as the real crawler.

How to block ChatGPT-User in robots.txt

We have not confirmed whether ChatGPT-User follows robots.txt. Add a rule, then check your logs to see that it stopped. robots.txt is a plain text file at the root of your domain, such as https://example.com/robots.txt.

Block ChatGPT-User from the whole site

User-agent: ChatGPT-User
Disallow: /

Block ChatGPT-User from some folders

User-agent: ChatGPT-User
Disallow: /private/
Disallow: /search

Allow ChatGPT-User only on some folders

User-agent: ChatGPT-User
Allow: /blog/
Disallow: /

Allow ChatGPT-User and block other crawlers

User-agent: ChatGPT-User
Allow: /

User-agent: *
Disallow: /

Things to know

  • Each host needs its own file. Rules on www.example.com do not apply to shop.example.com.
  • A crawler obeys only the most specific group that names it. Once you add a User-agent: ChatGPT-User group, ChatGPT-User ignores your User-agent: * rules, so repeat any that should still apply.
  • The user agent name is not case sensitive. We take ChatGPT-User from ChatGPT-User's user agent; check OpenAI's own documentation if it publishes a different robots.txt name.
  • Crawlers cache robots.txt, so changes can take up to a day to apply. Disallow stops crawling but does not remove pages already collected or indexed.

Limits of robots.txt

  1. ChatGPT-User may not follow the rule. robots.txt is a request that nothing enforces, and we have not confirmed that ChatGPT-User follows it. Check your logs after adding a rule.
  2. The user agent could be fake. robots.txt matches the name a request claims. Anyone can send ChatGPT-User's user agent, and a scraper using it is not bound by OpenAI's robots.txt policy. Only checking the source network tells a real ChatGPT-User from a fake one.

Treat robots.txt as a statement of your policy, and enforce it where requests arrive.

Who should care about ChatGPT-User

Publishers

ChatGPT-User reads pages for a person who asked an AI about them. The visit carries no ad impressions, so exclude it from ad requests and decide whether it may read paywalled content.

Transfon Security for publishers

Media buyers and advertisers

Agent visits can click through landing pages and fill in forms. Exclude ChatGPT-User from conversions so campaign optimisation is based on people.

Transfon Security for advertisers

E-commerce

Agents like ChatGPT-User can compare prices and may soon place orders. Decide whether agents may reach checkout, and verify them before they do.

Transfon Security for e-commerce

SaaS and APIs

ChatGPT-User can sign up, log in and call endpoints on a user's behalf. Rate limit it on sign-up and API routes, where fake trials and abuse start.

Transfon Security for SaaS and APIs

Block ChatGPT-User with Transfon Security

robots.txt is a request, not an enforcement. Transfon Security shows ChatGPT-User by name in the Top Bots report, so you can see how often it visits and which pages it reads. Block or allow it with a User Agent rule, or slow it down with a rate limit decision rule.

Transfon Security

Control ChatGPT-User on your site

  • See every ChatGPT-User visit and the pages it reads
  • Spot bots that fake trusted crawler names
  • Block, allow or rate limit ChatGPT-User with one rule
  • Works with or without robots.txt support

Frequently asked questions

What is ChatGPT-User?

ChatGPT-User is a bot in the AI assistants and agents category, operated by OpenAI. ChatGPT-User is for user actions in ChatGPT and Custom GPTs. When users ask ChatGPT or a CustomGPT a question, it may visit a web page to help answer and include a link to the source in its response.

Does ChatGPT-User respect robots.txt?

We have not confirmed whether ChatGPT-User follows robots.txt. Add a robots.txt rule, then check your logs to see that it stopped, and block it by user agent if it did not.

How do I block ChatGPT-User?

Add "User-agent: ChatGPT-User" followed by "Disallow: /" to the robots.txt file at the root of your site. To enforce it, block requests whose user agent contains "ChatGPT-User" at your WAF. In Transfon Security that is one User Agent block rule.

Is robots.txt enough to block ChatGPT-User?

No. robots.txt is voluntary, so ChatGPT-User may not follow it, and anyone can send ChatGPT-User's user agent, so a fake ChatGPT-User ignores your rule. Enforce the rule where requests arrive, at your WAF, and check the source network.

Should I block ChatGPT-User?

AI assistants fetch pages on demand when a user asks a chatbot or an agent to read, summarise or act on a URL. Each request is tied to a person, but the visit carries no ad impressions or analytics. Decide per agent whether the value is worth the load.

Related bots

All AI assistants and agents · Full bot directory

See Which Bots Visit Your Site

Transfon Security identifies every bot in this directory in your traffic reports, flags fake search crawlers, and lets you allow, block or rate limit each bot.

Stay ahead of the publishing curve

Privacy, security, monetization, and performance insights. A few emails a month, no spam.

By subscribing you agree to our privacy policy.