Perplexity AI: source access, citations, and visibility testing
Understand Perplexity's web-backed answers, crawler controls, citations, and changing search modes. Test source visibility and referral outcomes without unsupported ranking claims.
Perplexity AI
Perplexity describes itself as an AI-powered search engine that searches sources and generates conversational answers with citations. Its available models, modes, source sets, and interface change over time, so document the exact test conditions rather than treating “Perplexity” as one permanent answer system.
Web search is not the only possible context. Depending on the product and account, answers may also use selected files, organizational sources, connectors, or premium datasets. A citation shows attribution in a particular response; it does not prove that every claim is supported or that the cited page ranks consistently.
See Perplexity’s current product explanation for its own description of search and citations.
Crawler controls
Perplexity documents two relevant user agents with different purposes. PerplexityBot is intended to surface and link websites in Perplexity search results, while Perplexity-User supports user-initiated fetches and is not described as obeying robots.txt in the same way. Always check the current Perplexity crawler documentation, user-agent strings, and published IP guidance before changing access rules.
Allowing a crawler makes access possible; it does not guarantee indexing, selection, citation, or referral traffic. Confirm actual requests in server or CDN logs rather than relying only on a robots.txt tester.
Improve source readiness
- Return stable
200canonical pages with clear titles, authorship, dates, and primary content. - Make important claims precise and support them with methods, units, original evidence, and primary references.
- Keep essential content and links available in reliable HTML.
- Maintain accurate internal links and avoid redirect chains or inconsistent canonical signals.
- Publish corrections and meaningful updates without changing dates merely to look fresh.
Schema can support eligible search features elsewhere, but Perplexity does not document a universal schema recipe that earns citations. The same applies to “topical authority,” Reddit mentions, or any preferred word count.
Measure visibility reproducibly
Create a fixed prompt set by market, language, intent, mode, and account type. Save the full prompt, answer, model or mode, source filters, timestamp, cited URL, citation placement, and factual accuracy. Repeat the test and record misses as well as wins.
Separately measure verified bot access, citation presence, referral sessions from Perplexity domains, landing-page engagement, and downstream conversions. These are different stages; improvement in one does not prove improvement in another.