GPTBot crawls the web on its own schedule to collect training data. ChatGPT-User visits a page only when a ChatGPT user (or a custom GPT) asks it to — like a browser acting on someone’s behalf.
| GPTBot | ChatGPT-User | |
|---|---|---|
| Triggered by | OpenAI’s crawl schedule | A user’s request |
| Volume | Can be high | Proportional to user interest |
| robots.txt | Respected | May not apply to user-initiated fetches (see OpenAI docs) |
| Typical decision | Block if you opt out of training | Allow — these are real people’s questions |
Also see OAI-SearchBot vs GPTBot and the ChatGPT-User page.