Exa - Advanced Web Search
Exa is an MCP server that provides intelligent web search with superior ranking and filtering compared to standard search engines.
MCP Server Setup
The Exa MCP server should be configured in your MCP settings. The server URL is:
code
https://mcp.exa.ai/mcp
Configuration requires an Exa API key.
Available Tools
web_search
Advanced web search with intelligent ranking and filtering.
Parameters:
- •
query(string): Search query - •
numResults(number, optional): Number of results to return (default: 10) - •
includeDomains(array, optional): Restrict to specific domains - •
excludeDomains(array, optional): Exclude specific domains - •
startPublishedDate(string, optional): Filter by publish date (ISO format) - •
endPublishedDate(string, optional): Filter by publish date (ISO format)
Use cases:
- •Finding high-quality articles and sources
- •Research with domain-specific constraints
- •Time-sensitive queries (news, recent updates)
- •Comparing sources across specific websites
get_code_context
Deep search for code-related content across the web.
Parameters:
- •
query(string): Code-related search query - •
numResults(number, optional): Number of results - •
language(string, optional): Programming language filter
Use cases:
- •Finding code examples and snippets
- •Discovering implementations of specific algorithms
- •Researching libraries and frameworks
- •Finding documentation for coding problems
crawling_exa
Crawl and extract content from web pages.
Parameters:
- •
url(string): URL to crawl - •
depth(number, optional): Crawl depth (default: 1)
Use cases:
- •Extracting full content from specific URLs
- •Multi-page crawling with controlled depth
- •Content extraction for analysis
company_research_exa
Research companies and organizations.
Parameters:
- •
query(string): Company name or description - •
numResults(number, optional): Number of results
Use cases:
- •Company overviews and information
- •Finding articles about specific companies
- •Competitive research
- •Company news and updates
linkedin_search_exa
Search LinkedIn profiles and content.
Parameters:
- •
query(string): LinkedIn search query - •
numResults(number, optional): Number of results
Use cases:
- •Finding professionals and experts
- •Company employee research
- •Industry networking
- •Finding LinkedIn articles and posts
deep_researcher
Multi-step research across multiple queries and sources.
Parameters:
- •
query(string): Main research question - •
maxSteps(number, optional): Maximum research steps (default: 5)
Use cases:
- •Complex topics requiring multiple angles
- •Comprehensive research reports
- •Fact-checking across multiple sources
- •Building knowledge on unfamiliar topics
Best Practices
- •Prefer web_search for most queries - Exa's ranking is generally superior to standard search
- •Use domain filters - Restricting to trusted domains improves quality for technical queries
- •Leverage date filters - For news or time-sensitive information, always set date ranges
- •Use deep_researcher for complex topics - When you need comprehensive coverage from multiple angles
- •Combine tools - Use company_research_exa followed by web_search with specific domains for deeper insights
When Not to Use Exa
- •Simple, factual queries that don't require intelligent ranking
- •Searching your own local files or knowledge base
- •Real-time queries (Exa is not a live search engine)
- •Queries better suited to specialized APIs (e.g., weather, stock prices)