GEO Guide

GPTBot vs OAI-SearchBot: what's the difference?

OpenAI operates two separate web crawlers. Many site owners block one but not the other, or are unaware that both exist. Understanding the difference is important for full ChatGPT visibility.

GPTBot: the training crawler

GPTBot (user-agent: GPTBot) collects web content to train OpenAI's language models. It is the crawler that determines whether ChatGPT has knowledge of your site at training time. If GPTBot is blocked, your site cannot contribute to ChatGPT's knowledge base.

OAI-SearchBot: the real-time search crawler

OAI-SearchBot (user-agent: OAI-SearchBot) powers ChatGPT's web search feature. When a user asks ChatGPT to search the web, OAI-SearchBot fetches live pages to build the answer. If OAI-SearchBot is blocked, your site cannot appear in ChatGPT Search results.

You need to allow both

Allowing only GPTBot means ChatGPT knows about your site from training but cannot fetch it for live search results. Allowing only OAI-SearchBot means live search results may include your site, but ChatGPT's base knowledge lacks your content. Add both rules to get full coverage.

User-agent: GPTBot
Allow: /

User-agent: OAI-SearchBot
Allow: /

Free full report

Check your own site now.

Paste your URL and a target query. Get a citation readiness score, per-crawler breakdown, and copy-paste fixes in under 20 seconds.

Run a free scan →

Frequently asked

Can I allow OAI-SearchBot but block GPTBot?

Yes. Some publishers allow OAI-SearchBot (for live search citations) while blocking GPTBot (to prevent training data collection). OpenAI supports this distinction intentionally. However, for most sites that want maximum AI visibility, allowing both is the correct approach.