MCP Kagi Search

What is MCP Kagi Search?

MCP Kagi Search is an MCP server that provides the kagi_search tool, allowing AI assistants to perform web searches via the Kagi search engine. It is intended for developers who want to integrate Kagi search capabilities into MCP-compatible applications.

How to use MCP Kagi Search?

Install dependencies with uv sync, set the KAGI_API_TOKEN environment variable, run the server in development mode using mcp dev src/mcp_kagi_search/server.py, then test via the MCP Inspector at http://localhost:5173 using the kagi_search tool. Alternatively, install the CLI with uv pip install . and use the mcp-kagi-search command.

Key features of MCP Kagi Search

• Provides a single kagi_search tool for web searches.
• Designed for MCP-compatible AI assistants.
• Supports development mode with MCP Inspector.
• Includes a command-line interface (mcp-kagi-search).
• Requires a Kagi API token for authentication.

Use cases of MCP Kagi Search

• Integrate Kagi web search into AI chatbot conversations.
• Perform real-time research queries via an AI assistant.
• Build custom MCP workflows that include web search.
• Replace default search with Kagi in MCP applications.

FAQ from MCP Kagi Search

What environment variable is required?

Set KAGI_API_TOKEN to your Kagi API token.

How do I run the server for development?

Run mcp dev src/mcp_kagi_search/server.py and open the MCP Inspector at http://localhost:5173.

How do I test the tool?

In the MCP Inspector, run the kagi_search tool with a query (e.g., "steve jobs").

How do I install the CLI?

Install with uv pip install ., then verify with mcp-kagi-search --help.

What tool does this server provide?

It provides a single tool named kagi_search.