GPTBot

Operator

OpenAI

Type

AI Data Scraper

Official page

developers.openai.com

robots.txt token

GPTBot

Follows robots.txt

Yes, according to ai.robots.txt.

Published IP ranges

openai.com/gptbot.json

In Vireo

Excluded, reported under AI crawlers

What it does

Crawls content that may be used to train OpenAI's generative AI foundation models.

GPTBot is used to make our generative AI foundation models more useful and safe. It is used to crawl content that may be used in training our generative AI foundation models.

OpenAI

User agent

Requests from GPTBot carry a user agent like this:

Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; GPTBot/1.0; +https://openai.com/gptbot)

Device Detector identifies it with this pattern (a case-insensitive regular expression):

GPTBot

Anyone can put any name in a user agent. Where the operator publishes IP ranges, checking the request's address against them is the way to confirm a visit is genuine.

How to block or allow GPTBot

Add a group for its token to the robots.txt file at the root of your site. To keep it out of the whole site:

User-agent: GPTBot
Disallow: /

To let it crawl everything, use an empty Disallow: or Allow: / instead. WordPress serves a virtual robots.txt that many SEO plugins let you edit. A real robots.txt file uploaded to the site root takes its place.

Disallowing GPTBot indicates a site’s content should not be used in training generative AI foundation models.

OpenAI

robots.txt is a request, so it only works on crawlers that honour it. A bot that ignores it has to be blocked at the server or firewall, by user agent or IP address.

Worth knowing

Its setting is independent of OAI-SearchBot. OpenAI may reuse one crawl for both when a site allows both.

If your site has allowed both bots, we may use the results from just one crawl for both use cases to avoid duplicative crawling.

OpenAI

How Vireo counts GPTBot

Vireo keeps requests from GPTBot out of your pageviews and visitors and counts them on its bot report under AI crawlers, so you can see how much it left out rather than wonder why your server logs show more traffic.

Because it's an AI crawler, its requests also count toward the pages AI crawlers took, which Vireo shows next to the readers AI assistants sent back. See how the AI assistants card works.

Checked by running Vireo Analytics' own bot filter over GPTBot's user agents.

Sources

• OpenAI: crawler documentation (checked October 2026)
• GPTBot information page (platform.openai.com)
• Matomo Device Detector, bots.yml (LGPL-3.0)
• ai.robots.txt (MIT)