Tavily MCP
by Tavily
Real-time web search, extraction, mapping, and crawling for AI agents
search Node.js Beginner Self-hostable Verified
β 1.3k stars π
Updated: 1mo ago
Description
Production-ready MCP server providing real-time web search, intelligent data extraction, website mapping, and systematic web crawling. Optimized for AI agent research workflows with structured results and context-aware ranking. The search tool delivers fresh, relevant results from across the web, while extract pulls structured data from specific URLs. Web map discovers site structure for comprehensive coverage, and web crawl systematically processes multiple pages. Supports both local STDIO transport via npx and a hosted remote endpoint for zero-install access. Free tier available for getting started.
β Best for
AI agents that need real-time web information, structured extraction, and research capabilities
βοΈ Skip if
You need unlimited free search or already have a web search MCP that meets your needs
π‘ Use cases
- Real-time web research and fact-checking within AI assistant conversations
- Extracting structured data from specific web pages for analysis
- Mapping website structures to discover available content and endpoints
- Systematic web crawling for comprehensive data gathering
π Pros
- β Purpose-built for AI agent workflows with optimized result formatting
- β Free tier available β easy to get started without payment
- β Dual access via local npx or hosted remote MCP endpoint
- β Four complementary tools covering search, extract, map, and crawl
π Cons
- β Free tier has limited daily queries β paid plan needed for production use
- β Search results may vary in depth compared to traditional search engines
- β Crawling large sites can be slow and consume significant API quota
π§ Exposed tools (4 tools)
| Tool | Category | Description |
|---|---|---|
| web_crawl | crawling | Systematically crawl multiple pages of a website |
| tavily_extract | extraction | Extract structured data from a specific web page URL |
| web_map | mapping | Map the structure of a website to discover available pages |
| tavily_search | search | Search the web in real-time with AI-optimized results |
β‘ Installation
Prerequisites:
- β’ Node.js v18+
- β’ Tavily API key (free tier available)
- β’ API key required
Check Claude Code documentation to configure this MCP server.
π‘ Tips & tricks
Start with the free tier to evaluate. Use tavily_search for broad queries and tavily_extract when you need structured data from a known URL. The web_map tool is great for discovering all pages on a site before targeted extraction.
π Alternatives
Quick info
- Author
- Tavily
- License
- MIT
- Runtime
- Node.js 18+
- Transport
- stdio
- Category
- search
- Difficulty
- Beginner
- Self-hostable
- β
- Auth
- β
- Docker
- β
- Version
- latest
- Updated
- Feb 28, 2026
Client compatibility
- β Claude Code
- β Cursor
- β VS Code Copilot
- β Gemini CLI
- β Windsurf
- β Cline
- β JetBrains AI
- β Warp
Platforms
π macOS π§ Linux πͺ Windows