/
#52
Exa
Web search API built for LLMs and AI agents. Uses neural embeddings over a proprietary index for semantic retrieval instead of keyword matching. Offers multiple search modes from sub-250ms instant to deep-reasoning search, plus query-dependent highlights that condense full pages into dense, token-efficient excerpts for RAG. Custom indexes cover 1B+ people, 50M+ companies, and 100M+ research papers. Powers Cursor's @web feature.
Links
LAST 30 DAYS