Exa
Free tierWeb search API, website crawler, and deep research engine built for AI agents
Free tier available·Technical·API available
Key strengths
Fastest search API with results under 180ms (Exa Instant)Best-in-class accuracy on FRAMES, Tip-of-Tongue, and Seal0 benchmarksToken-efficient highlight extraction reducing token usage by up to 90%Supports structured outputs across 70M+ company enrichmentsUnified API covering search, crawling, and deep research in one endpoint
Free tier + paid plans
US
No ratings yet
Exa API — Developer Documentation
Authentication
All requests require an x-api-key header. Retrieve your key from the API Dashboard.
Core Endpoint: Search
POST https://api.exa.ai/search
Content-Type: application/json
x-api-key: YOUR_API_KEY
{
"query": "latest AI research on retrieval benchmarks",
"category": "research",
"num_results": 5,
"contents": {
"highlights": { "num_sentences": 3 }
}
}
Key Parameters
| Parameter | Type | Description |
|---|---|---|
query | string | Natural language or keyword search query |
category | enum | web, news, companies, research, people, financials |
num_results | int | Number of results to return |
contents.highlights | object | Extract relevant excerpts; reduces token usage by up to 90% |
livecrawl | boolean | Force real-time crawl of result pages |
Latency Modes
- Instant: <180ms — fastest available search provider
- Fast: ~450ms
- Auto: ~1s
- Deep: ~10s — agentic deep research with citations and structured outputs
Structured Outputs
Exa supports JSON-schema-based structured extraction, e.g., enriching company data (name, CEO, founded year) across 70M+ companies.
Additional Products
- Contents API: Retrieve full or highlighted page content from any URL
- Deep Agent API: Multi-step research pipeline returning grounded, cited answers
- MCP Server: Drop Exa into any MCP-compatible agent framework
- Websets: Programmable, persistent web data collections via API