indian-kanoon-mcp

# Indian Kanoon MCP Server An official Model Context Protocol (MCP) server for the Indian Kanoon API, built using the Python **FastMCP** framework. This server allows LLMs (like Claude, Cursor, ChatGPT, etc.) to query the Indian Kanoon legal database directly, search for court judgments, download document texts, view citations, and retrieve document fragments. --- ## Prerequisites & Installation The server is built to run seamlessly with **`uv`**, Astral's fast Python package installer and runner. ### 1. Install `uv` If you do not have `uv` installed, you can install it using one of the following methods: **macOS (Homebrew):** ```bash brew install uv ``` **Linux/macOS (curl):** ```bash curl -LsSf https://astral.sh/uv/install.sh | sh ``` **Windows:** ```powershell powershell -c "irm https://astral.sh/uv/install.ps1 | iex" ``` ### 2. Install Project Dependencies Navigate to the `mcp_server` directory and synchronize the dependencies: ```bash uv sync ``` ### 3. Set your API Token Ensure the `INDIANKANOON_API_TOKEN` environment variable is set. For local testing, you can export it in your terminal: ```bash export INDIANKANOON_API_TOKEN=your_actual_api_token_here ``` *(Alternatively, you can also place it in a `.env` file in the directory where you run the server).* --- ## How to Run & Test the Server Because this is a standard input/output (stdio) based MCP server, running it directly as a script (e.g. `uv run server.py`) will wait for JSON-RPC messages and throw a validation error if run interactively in a normal shell. Use the following methods to test and run it correctly: ### 1. Run the Integration Tests We have provided a test script to verify that your configuration and API token are working correctly: ```bash uv run python test.py ``` This runs a search query and fetches document fragments directly from the API. ### 2. Start the Interactive MCP Inspector FastMCP comes with a developer console/UI that lets you inspect and trigger the server's too…