What's the difference between GPTBot and OAI-SearchBot?
By the Pickgap team · Updated Oct 15, 2026 · 1 min read
GPTBot collects web content that may be used to train OpenAI's models. OAI-SearchBot finds and indexes pages so they can appear as sources in ChatGPT search. ChatGPT-User visits a page when a user asks ChatGPT to look at it. To be recommended in ChatGPT's answers, allow OAI-SearchBot; blocking GPTBot is a separate choice about training.
OpenAI's three user agents
| User agent | What it does | Block it and |
|---|---|---|
| GPTBot | Crawls content that may be used to train OpenAI's models | Your content is left out of future training |
| OAI-SearchBot | Finds and indexes pages for ChatGPT search results | Your site can't be shown as a source in ChatGPT search |
| ChatGPT-User | Visits a page when a user asks ChatGPT to open or read it | ChatGPT may fail when someone asks it to read your page |
They're controlled separately in robots.txt. Blocking one doesn't block the others.
What a local business should do
If you want ChatGPT to be able to cite your site, allow OAI-SearchBot. That's the one that matters for recommendations.
Whether to allow GPTBot is a matter of preference. Blocking it keeps your text out of future model training; it doesn't remove you from search answers. For most local businesses, being well known to AI does no harm, and many simply allow everything.
A quick check
Open yoursite.com/robots.txt and look for lines naming these agents. A rule like this blocks ChatGPT search:
User-agent: OAI-SearchBot
Disallow: /Also check your CDN or website builder for an "AI bots" setting, which can block OAI-SearchBot even when robots.txt allows it. The full list of AI crawlers, from all the major companies, is in robots.txt for AI crawlers.