In short. ChatGPT search uses web results in conversations, while OAI-SearchBot is the crawler site owners control for search access.
Definition: ChatGPT search is the web search capability within OpenAI's ChatGPT assistant. As of 17 October 2026, OpenAI's web search documentation describes search results and citations, references to sources, appearing in the conversation when web search is used.
Key facts
- OpenAI's web search guide says that people can request current information or sources in ChatGPT on the web.
- The same guide says workspace settings can limit search availability.
- OpenAI's crawler documentation identifies OAI-SearchBot as the crawler, software that automatically visits web pages, used to surface websites in ChatGPT search results.
- The crawler documentation gives site owners independent controls for OAI-SearchBot and GPTBot, the crawler associated with model training.
How does search appear in a conversation?
OpenAI's web search guide describes a conversational workflow: someone asks for current information or sources, and search results and citations appear in the chat when search runs. In the desktop app, search activity is recorded alongside other tool calls in the conversation transcript.
The guide instructs readers to treat web results as untrusted input. For a business checking an answer about its services, this makes the linked source material part of the review: the relevant page can be read alongside the answer's claim.
Which website controls affect search access?
A site's robots.txt file communicates crawling rules to such software. OpenAI's crawler overview recommends permitting OAI-SearchBot and requests from its published IP ranges, meaning the network addresses used by the crawler.
The following example expresses the separate search and training choices described in that overview. It allows search crawling and disallows GPTBot crawling:
User-agent: OAI-SearchBot
Allow: /
User-agent: GPTBot
Disallow: /
OpenAI's documentation says a search system adjustment can take approximately 24 hours after a robots.txt change. It also says sites opting out of OAI-SearchBot can still appear as navigational links, which direct someone to a website.
What should a website owner check?
The practical access check follows OpenAI's crawler documentation: inspect the OAI-SearchBot rules and check whether requests from the published crawler IP ranges are permitted. These checks address the access settings OpenAI recommends.
The same document identifies ChatGPT-User as an agent for certain user-initiated page visits. Because a user initiates those actions, robots.txt rules may not apply. Search opt-outs and automatic crawling are managed through OAI-SearchBot.
Related terms
FAQ
Can you allow search crawling while declining training crawling?
Yes. OpenAI's crawler overview describes independent settings for OAI-SearchBot and GPTBot. The example above expresses that choice.
Where can you find evidence that a search ran?
OpenAI's web search guide says search results and citations appear in the web chat. In the desktop app, you can review search activity in the conversation transcript.
Sources
Check how a page reads to answer engines. Open the GEO score analyzer