Discover and explore top open-source AI tools and projects—updated daily.
kagisearchAI tool integration server for Kagi search and summarization
Top 97.9% on SourcePulse
Summary
This project provides the official Model Context Protocol (MCP) server for Kagi Search, enabling integration with AI tools. It allows users to leverage Kagi's search and summarization capabilities within LLM-powered applications.
How It Works
The kagimcp project implements the Model Context Protocol to expose Kagi's search API and summarization models as a server for AI tools like Claude and Codex CLI. It utilizes the uv package manager for dependencies and uvx for command execution, simplifying setup and ensuring consistent environments. This MCP architecture promotes interoperability.
Quick Start & Requirements
uv (via provided script), then run uvx kagimcp. Alternatively, use npx -y @smithery/cli install kagimcp --client claude.support@kagi.com) and an API key (KAGI_API_KEY).codex mcp add kagi --env KAGI_API_KEY=<YOUR_API_KEY_HERE> -- uvx kagimcpclaude_desktop_config.json or via claude mcp add command, specifying KAGI_API_KEY and optionally KAGI_SUMMARIZER_ENGINE. Claude's native web search may need disabling.uv, create uv venv, uv sync. Debugging via mcp dev or npx @modelcontextprotocol/inspector.support@kagi.com.Highlighted Details
KAGI_SUMMARIZER_ENGINE environment variable (e.g., "cecil", "daphne").FASTMCP_LOG_LEVEL environment variable.--http) with configurable host/port.Maintenance & Community
https://github.com/kagisearch/kagimcp/issues/4.Licensing & Compatibility
Limitations & Caveats
uv, codex, claude), which may add to the initial learning curve.1 month ago
Inactive