AI search & GEO

GPTBot

Also called: OpenAI GPTBot, GPTBot crawler

GPTBot is OpenAI's web crawler that collects publicly accessible content to train its GPT foundation models. Site owners allow or block it in robots.txt via the user-agent token "GPTBot," which is separate from OpenAI's search crawler (OAI-SearchBot) and its live user-fetch agent (ChatGPT-User).

GPTBot first appeared in August 2023. Its full user-agent string looks like Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.4; +https://openai.com/gptbot (the version number changes over time). OpenAI publishes GPTBot’s IP ranges at openai.com/gptbot.json, so you can verify a request is genuine.

GPTBot is not OpenAI’s only crawler

It is one of three separate agents, each with its own job:

Disallowing GPTBot in robots.txt removes your content from future training runs. It does not remove you from ChatGPT search results, because those are governed by OAI-SearchBot. Each agent needs its own rule:

User-agent: GPTBot
Disallow: /

One 2025 change matters here: OpenAI now states that ChatGPT-User may ignore robots.txt because those fetches are user-initiated, so robots.txt reliably controls only GPTBot and OAI-SearchBot. That means the block decision that most affects your AI visibility is the search crawler, not the training one.

How it affects your traffic

For AI-search visibility, GPTBot is the wrong bot to obsess over. What gets you cited in ChatGPT is OAI-SearchBot and ChatGPT-User, so a common self-inflicted wound is blanket-blocking every OpenAI agent (usually from a copy-pasted robots.txt) and quietly deleting yourself from ChatGPT answers while believing you only opted out of training. Our AI SEO service audits these crawler rules so you can protect training content if you choose to, without severing the crawl paths that actually drive AI referral traffic.

Get AI SEO that moves the needle

We turn terms like this into ranked pages and qualified pipeline. Start with a free Initial SEO Strategy.