==================
/// MCP ///
/// KAG ///
==================
[server:online]
[protocol:ready]kagimcp
by kagisearch
Official Model Context Protocol (MCP) server that exposes Kagi web-search, summarizer and related tools to MCP-compatible clients (Claude Desktop/Code, etc.).
145
18
Enterprise01
search
Perform a web search using Kagi’s Search API
02
summarizer
Generate summaries of content (e.g., YouTube videos) using Kagi’s Summarizer engine
Installation
1. Clone the repository:
git clone https://github.com/kagisearch/kagimcp.git && cd kagimcp2. Create and activate a Python virtual environment (recommended):
python3 -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate3. Install server dependencies:
pip install -r requirements.txt
# If a development install is required:
# pip install -r requirements-dev.txt4. (Optional) Create an .env file or export environment variables required by MCP, for example:
export MCP_API_KEY="<your-secret-key>"
export MCP_PORT=80005. Launch the MCP server locally:
python -m kagimcp
# or, if provided, uvicorn kagimcp.app:app --host 0.0.0.0 --port ${MCP_PORT:-8000}6. Verify the server is running:
curl http://localhost:8000/health
# Expected response → {"status":"ok"}7. (Container alternative) Build and run via Docker if a Dockerfile is present:
docker build -t kagimcp .
docker run -p 8000:8000 -e MCP_API_KEY=<key> kagimcpDocumentation
License: MIT License
Updated 7/30/2025