mcp-askcpg

Qmed-AskCPG is a Python-based MCP server that lets clinicians and developers query Malaysian Clinical Practice Guidelines (CPGs) using natural language. Once the MCP server is running, you can ask questions in plain language such as “What is the recommended procedure for stroke management in Malaysia’s CPGs?” The server processes the query, searches official Malaysian CPG references, and returns a structured, citation-backed response that highlights the relevant guidelines and sections. This makes it easier to extract evidence-based recommendations directly from authoritative sources.

To operate, you’ll typically start the MCP server with a container or task runner (as shown in the installation guidance). You then send natural language prompts to the server; the MCP layer handles parsing, retrieving, and formatting the answer with references. Be sure to provide any required API keys or backend configuration via environment variables if the MCP setup requires external services for fetching or validating CPG content.

Prerequisites:

• Python 3.8+ (for the mcp-askcpg package)
• pip (Python package installer)
• Access to the network to fetch or validate CPG references (if required by backend)

Installation steps:

1. Ensure Python and pip are installed on your system.

2. Install the mcp-askcpg package from PyPI:

   pip install mcp-askcpg

3. Verify installation by running the module or using the provided task runner configuration (uvx in this example).

4. Prepare environment variables if your backend requires authentication or specific endpoints:

   • ASKCPG_API_KEY: Your API key for the ASKCPG backend (if applicable)
   • ASKCPG_BACKEND: The backend URL or identifier to connect to (if applicable)

5. Start the MCP server using the configuration example provided in the documentation or your task runner:

   uvx mcp-askcpg

If you prefer a different runner or a programmatic setup, adapt the command according to your environment and the mcp-askcpg package's instructions.

Tips and common issues:

• Ensure that any required API keys or backend endpoints are correctly set in the environment variables (ASKCPG_API_KEY, ASKCPG_BACKEND).
• If you encounter network or authentication errors, double-check access permissions and that the API keys are valid.
• The MCP server returns structured responses with references to official CPG documents; if references are missing, verify backend connectivity and the data sources configured for the ASKCPG backend.
• If you plan to deploy in production, consider securing environment variables and using a container orchestration platform to manage the MCP service lifecycle.
• This setup assumes the uvx-based runner for configuration; if you use a different runner, adjust the mcp_config accordingly (e.g., npx, docker, or node-based setups).
• For updates, keep the mcp-askcpg package in sync with the MCP framework version you’re using to avoid compatibility issues.