home / mcp / steampipe mcp server

Steampipe MCP Server

This repo contains an MCP written in Python for Steampipe.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "b0ttle-neck-mcp-steampipe": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/inspector",
        "uv",
        "--directory",
        ".",
        "run",
        "steampipe_mcp_server.py"
      ],
      "env": {
        "STEAMPIPE_CONFIG": "~/.steampipe/config/github.spc"
      }
    }
  }
}

This MCP server acts as a bridge between your AI model and the Steampipe tool, enabling you to query Steampipe data sources via an MCP client. It provides a lightweight interception layer that lets your model send queries and receive results through a standardized interface.

How to use

To use this MCP server, you run the MCP Inspector to test and validate the integration, then connect your MCP client to the server and start issuing queries against Steampipe data sources.

How to install

Prerequisites you need to have installed before starting are listed below. Follow the concrete steps to set up the environment and run the server.

Install Python 3.10 or newer, then install uv and the MCP CLI. Ensure Steampipe is installed and configured with at least one plugin (for example, the GitHub plugin) and credentials available in the Steampipe config directory.

Install Node.js and npx, which are required for the MCP Inspector and may be needed for additional MCP tooling.

# Example setup steps
# 1) Install Python 3.10+
# 2) Install uv and MCP CLI
pip install uvx
pip install mcp[cli]
# 3) Install Steampipe and configure plugins
# See Steampipe installation instructions for your platform
# 4) Install Node.js and npx
# 5) Prepare the MCP Inspector and verify tooling

Run the MCP Inspector to test the server

The Inspector helps you confirm that your MCP server is wired correctly and can interact with Steampipe. Run the following command to start the Inspector and point it at your MCP server script.

npx -y @modelcontextprotocol/inspector uv --directory . run steampipe_mcp_server.py

Available tools

run_steampipe_query

Tool to execute a Steampipe query against the configured sources via the MCP server and return results to the MCP client.