How does ChatGPT choose which sites to cite?
ChatGPT prioritizes citing pages with clear answers, structured data, and trust signals when web search is enabled. The page must be crawlable and current; login walls and crawler blocks make it ineligible.
Updated 2026-07-11
Frequently asked questions
- How does ChatGPT choose which sites to cite?
- ChatGPT prioritizes citing pages with clear answers, structured data, and trust signals when web search is enabled. The page must be crawlable and current; login walls and crawler blocks make it ineligible.
- What's the difference between OAI-SearchBot and ChatGPT-User?
- OAI-SearchBot indexes pages ahead of time so they can become citation candidates in search-based answers, while ChatGPT-User is a separate agent that only visits a page in real time when a user asks ChatGPT to open a specific link during a conversation. The two are controlled independently in robots.txt, so at minimum OAI-SearchBot needs to stay allowed if you want a shot at being cited.
- Does the same logic apply to Perplexity and Google?
- Yes. Perplexity is inherently search-based, and Google's AI Overview runs a similar screening. Any search-driven AI relies on the same three things: crawlability, clarity, and trust signals.