Overview
What is Tavily Search MCP Server?
Tavily Search MCP Server is an MCP server that integrates the Tavily Search API, providing optimized web search capabilities for large language models. It is designed for developers and AI applications that need structured, relevant search results.
How to use Tavily Search MCP Server?
To use it, clone the repository, install dependencies with npm install, and build with npm run build. Then configure the server in Claude Desktop by adding a tavily-search-server entry to the mcpServers object in claude_desktop_config.json, providing your Tavily API key. You can run the server via Node.js (node dist/index.js) or Docker.
Key features of Tavily Search MCP Server
- Web search optimized for LLMs with depth and topic control.
- Content extraction for high-quality, relevant snippets.
- Optional inclusion of images, image descriptions, and LLM answers.
- Optional raw HTML content retrieval.
- Domain inclusion and exclusion filtering.
Use cases of Tavily Search MCP Server
- Enabling an LLM to retrieve up-to-date information from the web.
- Performing news searches with configurable time ranges.
- Filtering search results to or from specific domains.
- Extracting summarized answers for user queries.
FAQ from Tavily Search MCP Server
What transport options are supported?
The server can be run over stdio (default) or SSE (by running node dist/sse.js and specifying a port).
How do I get a Tavily API key?
Sign up for a Tavily API account, choose a plan (free tier available), and generate your API key from the Tavily dashboard.
Can I run the server using Docker?
Yes, build the Docker image with docker build -t tavily-search-mcp-server:latest . and run it with docker run -it --rm -e TAVILY_API_KEY="your_key" tavily-search-mcp-server:latest. For SSE transport, add -e TRANSPORT="sse" and map port 3001.
Where do I configure the server for Claude Desktop?
Edit your Claude Desktop configuration file (~/Library/Application\ Support/Claude/claude_desktop_config.json on macOS or %APPDATA%\Claude\claude_desktop_config.json on Windows) and add the server’s command and environment variables under mcpServers.
What are the runtime dependencies?
The server requires Node.js to build and run. For Docker usage, Docker must be installed. No other external services are needed beyond the Tavily API.
