| name | exa-search |
| description | Neural web search, code search, and deep research via Exa AI. Use for semantic web queries, finding code examples, company research, and multi-source synthesis. |
Exa Search
Neural web search, code search, and deep research via Exa AI.
When to Use
- Semantic web search for current information
- Finding code examples and documentation
- Company research and competitive analysis
- Deep research tasks requiring multi-source synthesis
- Finding similar content to a given URL
Tools
| Tool | Description |
|---|
exa_web_search_exa | Search the web for any topic |
exa_get_code_context_exa | Find code examples, docs, programming solutions |
exa_crawling_exa | Get full content of a specific URL |
exa_company_research_exa | Research a company's info, news, insights |
exa_linkedin_search_exa | ⚠️ DEPRECATED — use people_search_exa |
exa_deep_researcher_start | Start AI research agent for detailed reports |
exa_deep_researcher_check | Check status/results of deep research task |
Usage Patterns
Web Search
exa_web_search_exa query="Effect-TS Schema.TaggedStruct patterns" numResults=5
Code Search
exa_get_code_context_exa query="effect-atom Atom.runtime React hook pattern"
Crawl a URL
exa_crawling_exa url="https://effect.website/docs/schema/basic-usage"
Deep Research (async)
1. exa_deep_researcher_start instructions="Compare Effect-TS vs Zod schema validation"
2. exa_deep_researcher_check taskId="<id from step 1>"
Best Practices
- Code search — Use
get_code_context_exa for programming questions
- Web search — Use
web_search_exa for general knowledge
- Deep research is async — Start then check status after a delay
- numResults — Default 8, lower for focused queries
