What each provider documents
OpenAI says ChatGPT can choose to search the web and can also search when a user selects that option. Search responses include links to web sources. OpenAI documents OAI-SearchBot as the crawler used to surface websites in ChatGPT search.
Perplexity documents a Search API that returns ranked web results and a separate Sonar API for cited answers. Its crawler documentation identifies PerplexityBot as the search crawler used to surface and link websites in Perplexity results.
What cannot be inferred from public documentation
- There is no documented fixed number of sources that every response must cite.
- Neither provider publishes a complete source-selection or citation-ranking formula.
- A citation observed for one prompt, account, date or locale is not proof of stable visibility.
- A crawler allow rule creates technical eligibility, not a citation guarantee.
Run a defensible comparison
- Freeze identical prompts, language, location, date and account conditions where the interface exposes them.
- Record whether the brand is mentioned, whether a link is present, whether attribution is accurate and which sources support the answer.
- Repeat the same panel over time before reporting a trend.
- Use server logs and indexing evidence to diagnose discovery separately from answer selection.