Exa
Neural semantic search engine built for AI — agents search by meaning, not keywords, across the entire web.
About
Exa is a search engine built from the ground up for AI agents. Unlike traditional keyword-based search, Exa uses neural embeddings to find results by meaning — so an agent searching for "tools that help autonomous agents manage OAuth flows" gets relevant results even if those exact words don't appear on the page. The API returns full page content, not just snippets and links, making it immediately useful for agent workflows.
Beyond basic search, Exa offers specialized modes for finding people (LinkedIn profiles), companies, code (GitHub, Stack Overflow, docs), and research papers. The Research endpoint performs autonomous multi-step deep research, and Websets can collect structured datasets from across the web. Agents connect via API key or through Exa's official MCP server, which exposes tools like web_search_exa, company_research_exa, and get_code_context_exa.
The free tier includes 1,000 requests per month, with paid plans starting at $7 per 1,000 searches. Exa is SOC 2 Type II certified and offers zero data retention for enterprise deployments.