kagi-server MCP Server

What is kagi-server?

A TypeScript-based MCP server that integrates the Kagi Search API, providing tools for web search operations. It is designed for use with AI assistants like Claude Desktop.

How to use kagi-server?

Install dependencies with npm install, build with npm run build, set the KAGI_API_KEY environment variable in a .env file, and add the server configuration to Claude Desktop's config file (e.g., claude_desktop_config.json). Alternatively, install via Smithery using npx @smithery/cli install kagi-server --client claude.

Features of kagi-server

• kagi_search tool for web searches via Kagi API
• Takes a query string and optional limit parameter
• Built with TypeScript and the MCP SDK
• Installable via Smithery or manual config
• Debugging support via MCP Inspector

Use cases of kagi-server

• Search the web for current information through Claude
• Retrieve and analyze search results with an AI assistant

FAQ from kagi-server

What tools does kagi-server implement?

Only kagi_search is currently implemented. kagi_summarize, kagi_fastgpt, and kagi_enrich are planned but not yet available.

How do I get a Kagi API key?

Create a .env file in the project root with KAGI_API_KEY=your_api_key_here. The Kagi API is currently in private beta.

How do I debug the server?

Use the MCP Inspector by running npm run inspector, which provides a URL for debugging tools in a browser.

How do I install kagi-server for Claude Desktop?

Install via Smithery (npx @smithery/cli install kagi-server --client claude) or manually by adding the server config to claude_desktop_config.json with the path to build/index.js and the KAGI_API_KEY environment variable.

What transport does the server use?

The server communicates over stdio using the Model Context Protocol.