physionet-mcp

Lean MCP server for PhysioNet datasets - works with any PhysioNet dataset you have access to.

📺 This is a lean version of m3 with similar BigQuery and PhysioNet setup. Check out detailed videos here: https://rafiattrach.github.io/m3/

Install uv (required for uvx)

We use uvx to run the MCP server. Install uv from the official installer, then verify with uv --version.

• macOS:

brew install uv

• Linux (or macOS without Homebrew):

curl -LsSf https://astral.sh/uv/install.sh | sh
# macOS - enable for GUI apps like Claude Desktop:
sudo ln -s $(which uv) $(which uvx) /usr/local/bin/

• Windows (PowerShell)

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Verify installation:

uv --version

BigQuery authentication (CLI)

1. Install Google Cloud SDK:
   • macOS (Homebrew): brew install google-cloud-sdk
   • Windows/Linux: see the installer at https://cloud.google.com/sdk/docs/install
2. Authenticate Application Default Credentials (ADC):

gcloud auth application-default login

This will open your browser — choose the Google account that has access to your BigQuery project with PhysioNet data.

3. Use your Google Cloud project ID in the MCP config (see Quick Setup). You can also export it in your shell:

export BIGQUERY_PROJECT_ID=your-project-id

Quick Setup

Paste the following into your MCP client configuration, then restart your client.

Production

{
  "mcpServers": {
    "physionet-mcp": {
      "command": "uvx",
      "args": ["physionet-mcp"],
      "env": {
        "BIGQUERY_PROJECT_ID": "your-project-id"
      }
    }
  }
}

Local Development

{
  "mcpServers": {
    "physionet-mcp": {
      "command": "/path/to/physionet-mcp/venv/bin/python",
      "args": ["-m", "physionet_mcp.mcp_server"],
      "cwd": "/path/to/physionet-mcp",
      "env": {
        "BIGQUERY_PROJECT_ID": "your-project-id"
      }
    }
  }
}

Replace your-project-id with your Google Cloud project ID.

4 Simple Tools

1. list_accessible_datasets → See what you can access
2. get_database_schema → Find tables in a dataset
3. get_table_info → Check structure & sample data
4. execute_query → Run your analysis

Usage Examples

• “What PhysioNet datasets can I access?”
• “Show me MIMIC-IV hospital tables”
• “What’s in the patients table?”
• “How many patients are in MIMIC-IV?”

Future Enhancements

Potential improvements for enterprise use:

• Dataset filtering - Restrict access to specific datasets for security
• Query optimization - Add result caching and query cost tracking
• Rate limiting - Implement query throttling for shared environments
• Enhanced metadata - Add column descriptions and data quality metrics

License

MIT