robots.txt rules
Copy the group you need into the robots.txt file at the root of your site. Keep only one option per bot.
At a glance
| Detail | GPTBot |
|---|---|
| Operator | OpenAI |
| Type | Training crawler |
| What it does | Collects public web pages that may be used to train OpenAI's models. |
| robots.txt token | GPTBot |
| Example user agent | Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.3; +https://openai.com/gptbot |
| Follows robots.txt | OpenAI says GPTBot follows robots.txt. |
| How to verify | Check the request IP against the list OpenAI publishes at openai.com/gptbot.json. |
| If you block it | Your future content is not collected for OpenAI model training. ChatGPT search is not affected, because it uses OAI-SearchBot. |
| Source | OpenAI crawler documentation (developers.openai.com) |
User agent version numbers change over time. Match on the token, not the full string.
What GPTBot does
GPTBot is OpenAI's training crawler. It visits public pages to gather text that may be used to train future GPT models. It is separate from the crawler that powers ChatGPT search, so it has nothing to do with whether ChatGPT can find and cite your pages today.
Should you allow or block it?
Your choice. Blocking it is a training decision only. Allowing it means future models are more likely to know your brand from training.
If you block it: Your future content is not collected for OpenAI model training. ChatGPT search is not affected, because it uses OAI-SearchBot.
- Training crawlers decide whether your content helps build future models. Blocking them does not remove you from AI search today.
- Search crawlers decide whether AI engines can find and cite your pages. Blocking them usually costs visibility.
- User triggered fetchers read a page because a person asked. Blocking them stops the AI reading you at the moment someone is interested.
How to find GPTBot in your logs
Search your server or CDN access logs for the token GPTBot. This command counts visits by IP, path and status code in a standard access log:
grep -i "GPTBot" access.log | awk '{print $1, $7, $9}' | sort | uniq -c | sort -rn | head
A run of 403 or 503 status codes means your firewall or bot protection is blocking it, even if robots.txt allows it. That is one of the most common hidden reasons sites lose AI visibility.
How to tell if a visit is genuine
Anyone can put GPTBot in a user agent string, so the name alone proves nothing. Check the request IP against the list OpenAI publishes at openai.com/gptbot.json.
Other OpenAI bots
- OAI-SearchBot: ai search crawler. Finds and indexes pages so they can appear and be cited in ChatGPT search.
- ChatGPT-User: user triggered fetcher. Opens a page when a ChatGPT user, custom GPT or agent asks it to.