What is GPTBot?
GPTBot is one of the AI crawlers OpenAI runs. OpenAI says it "is used to crawl content that may be used in training our generative AI foundation models" (OpenAI: Overview of OpenAI crawlers, read 1 November 2026).
It identifies itself in the user agent string. OpenAI gives this example, and notes the version number may change:
- Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.4; +https://openai.com/gptbot
OpenAI publishes the IP addresses GPTBot uses at openai.com/gptbot.json. Use that list to check that a visit claiming to be GPTBot really comes from OpenAI.
Does blocking GPTBot remove you from ChatGPT?
No, not from ChatGPT search. OpenAI runs a separate crawler for that, OAI-SearchBot. OpenAI writes that "each setting is independent of the others". A site can allow OAI-SearchBot to appear in search results and disallow GPTBot to keep its content out of training.
OpenAI lists two more agents. ChatGPT-User visits a page when a person asks ChatGPT something. It is not an automatic crawler, and OpenAI says robots.txt rules may not apply to it. OAI-AdsBot only visits pages submitted as ads in ChatGPT, and OpenAI says its data is not used to train foundation models.
Why does GPTBot matter for B2B SaaS?
The choice is a trade-off, and it is yours. Some teams block GPTBot to keep their content out of training. Others allow it, because they want future models to know their product and category.
Two things are certain from OpenAI's own page. Blocking GPTBot does not change how OAI-SearchBot treats you. And OpenAI may use one crawl for both purposes when you allow both bots, "to avoid duplicative crawling".
What is not certain: how much training data shapes whether a model names your brand later. No source we read measures that. Treat any claim that blocking or allowing GPTBot changes AI mentions with care.
How do you allow or block GPTBot?
You add a group for GPTBot to your robots.txt file at the root of your domain. A hypothetical example that blocks training but keeps ChatGPT search open:
- User-agent: GPTBot
- Disallow: /
- User-agent: OAI-SearchBot
- Allow: /
- Open yourdomain.com/robots.txt and check what is there now. Check your CDN or firewall settings too, since they can block bots before robots.txt matters.
- Decide per bot: training (GPTBot), search (OAI-SearchBot).
- Add or change the groups and publish the file.
- Check your server logs for the GPTBot user agent and compare the IPs with openai.com/gptbot.json.
What does this look like in practice? (hypothetical example)
A hypothetical HR software company blocked every bot with "AI" or "GPT" in its name in 2024. In 2026 it wants to show up in ChatGPT search answers. It reads OpenAI's page, sees that search runs on OAI-SearchBot, and opens that bot while keeping GPTBot blocked. The two choices no longer get mixed up.
How does GPTBot relate to third-party lists?
Your robots.txt only controls your own site. GPTBot can still read the review sites, comparison pages and lists that describe your product, under their rules. So the pages other people write about you matter whatever you decide. The LLM SEO guide covers how brands get into those pages.