VectorCiteVectorCite
Perplexity AI

Perplexity

Pure RAG answer engine — every response is grounded in cited sources.

How it works

Perplexity is purpose-built as an answer engine: every query triggers retrieval (no 'should I search?' decision). Default model is Sonar. Retrieval uses hybrid search (dense + BM25) over Perplexity's own crawled index plus a partnership-feed (Yelp, Brave, others). Sources are reranked and the top 5 are synthesized into a cited paragraph. Perplexity's Query Fanout feature decomposes complex queries into 3–5 sub-queries and aggregates citations across all — a behavior other engines now copy.

Official docs
User-agents
PerplexityBotPerplexity-User

Allow these in your robots.txt explicitly. Audit your current setup with the robots.txt analyzer.

Signals it weights heavily
  • Query coverage + sub-query coverage
  • Citation density
  • Recency
  • Structural clarity (H2 + lists)
  • BLUF answers
How to optimize for Perplexity
  • Allow PerplexityBot explicitly in robots.txt.
  • Optimize for sub-query coverage — your page should answer the seed query AND its decomposed sub-queries.
  • Direct, factual style works best — Perplexity's reranker prefers concise direct answers over rhetorical setups.
  • Recency boost — re-publish evergreen pillar posts with updated dates every 6 months.

See your score on Perplexity's signals.

Run free audit

Other engines