Exa
Search engine and API built for AI agents, offering semantic web search, content retrieval, and agentic research tools.
Last verified:
Exa (formerly Metaphor) is an applied AI lab building a search engine designed specifically for AI applications and agents. Its API provides semantic web search, page content/highlights extraction, an autonomous research agent, and a deep-search product with structured outputs. Exa is used by companies including Cursor, Cognition, and HubSpot to ground LLMs and agents with fresh web knowledge.
Pricing
| Tier | Price | Key features |
|---|---|---|
| Free | $0/monthly | Web search tool calls for agents, Webpage text/highlights, Configurable latency (180ms-1s) |
| Search | $7/usage | Semantic web search, Up to 10 results per request |
| Agent | Custom | Autonomous research agent, Configurable effort modes, $0.10 per Agent Compute Unit, $0.005 per search tool call |
| Deep Search | $12/usage | Structured outputs, Deeper multi-step search |
| Enterprise | Custom | Custom rate limits (QPS), Volume discounts, Custom pricing |
Source: https://exa.ai/pricing
Agent-readiness
- Discoverability
- 78/100
- Machine-readable content
- 82/100
- Programmatic access
- 90/100
- MCP / agent support
- 88/100
- Self-serve transactability
- 82/100
- Transparency
- 62/100
Security & compliance
- SSO / SAML
- Not publicly documented
- Encryption at rest
- Not publicly documented
- Encryption in transit
- Yes
- Audit log
- Not publicly documented
- Pentest report
- Not publicly documented
"Not publicly documented" means Crail found no public evidence either way — not a confirmed absence. See our methodology. Trust center: https://trust.exa.ai/.
FAQ
Does Exa have a free tier?
Yes, Exa offers a free tier or free trial.
Can you buy Exa without talking to sales?
Yes, Exa can be purchased self-serve without a sales call.
Does Exa have an official MCP (Model Context Protocol) server?
Yes, Exa publishes an official MCP server (https://github.com/exa-labs/exa-mcp-server).
Reviews
No reviews yet for Exa. Be the first to submit a verified review.
Methodology & disclosure
Data verified via agent crawled, last checked 2026-07-27. Agent-readiness methodology version crail-ar-v0.1. No vendor payment influences these scores — see our methodology page.