This project provides a server implementation for the Model Context Protocol (MCP), enabling AI models to interact with Tavily's web search and data extraction tools. It targets developers and users of MCP-compatible AI clients like VS Code, Cursor, and Claude Desktop, offering real-time web data access to enhance AI capabilities.

How It Works

The Tavily MCP server acts as a bridge, translating requests from MCP clients into calls to the Tavily API. It leverages Node.js and npx for execution, facilitating seamless integration with tools like tavily-search and tavily-extract. This approach allows AI models to access up-to-date web information and extract structured data from web pages directly within their workflow.

Quick Start & Requirements

• Install/Run: npx -y tavily-mcp@0.1.4
• Prerequisites: Tavily API key, Node.js (v20+), Git.
• Setup: Minimal, primarily configuration within MCP clients.
• Docs: Tavily MCP Server

Highlighted Details

• Seamless integration with VS Code, Cursor, and Claude Desktop.
• Supports tavily-search for real-time web search and tavily-extract for data extraction.
• Provides tutorials for integrating with Neo4j and Cline within VS Code.

Maintenance & Community

• Developed by Tavily AI.
• Links to configuration guides for various clients are provided.

Licensing & Compatibility

• License not explicitly stated in the README.
• Compatible with MCP clients; commercial use depends on Tavily's API terms.

Limitations & Caveats

The README mentions a specific version (0.1.2) for Claude Desktop configuration, which might differ from the latest npx version. Troubleshooting tips are provided for common issues like server not found or API key problems.