home / mcp / conductor mcp server

Conductor MCP Server

Provides tools to interact with a Conductor instance for workflow creation, execution, and analysis.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "conductor-oss-conductor-mcp": {
      "command": "conductor-mcp",
      "args": [
        "--config",
        "<ABSOLUTE PATH TO A JSON CONFIG FILE>"
      ],
      "env": {
        "CONDUCTOR_AUTH_KEY": "YOUR_APPLICATION_AUTH_KEY",
        "CONDUCTOR_SERVER_URL": "https://developer.orkescloud.com/api",
        "CONDUCTOR_AUTH_SECRET": "YOUR_APPLICATION_SECRET_KEY"
      }
    }
  }
}

You can run the Conductor MCP Server to interact with a Conductor instance for creating, running, and analyzing workflows. This server acts as a bridge between your AI agents and Conductor, enabling you to manage workflows through MCP client tooling.

How to use

Install the MCP server client, configure it with your Conductor credentials, and run the server locally so your AI assistant can create and manage Conductor workflows. You will provide a JSON config with your Conductor server URL and authentication keys, then point your MCP-compatible agent at the local server command to start using it.

How to install

Prerequisites: Python and pip must be available on your system.

Step 1: Install the Conductor MCP package from PyPI.

pip install conductor-mcp

Additional configuration and running the server

Prepare your Conductor connection by creating a JSON config with your Conductor server URL and authentication keys.

{
  "CONDUCTOR_SERVER_URL": "https://developer.orkescloud.com/api",
  "CONDUCTOR_AUTH_KEY": "<YOUR_APPLICATION_AUTH_KEY>",
  "CONDUCTOR_AUTH_SECRET": "<YOUR_APPLICATION_SECRET_KEY>"
}

Start the MCP server locally

Use the following runtime command to launch the MCP server and load your JSON config. The command shown starts the MCP server and passes your absolute path to the config file.

{
  "mcpServers": {
    "conductor": {
      "command": "conductor-mcp",
      "args": [
        "--config",
        "<ABSOLUTE PATH TO A JSON CONFIG FILE>"
      ]
    }
  }
}

Alternative startup method with a local runner

If you are using the local runner flow, you can run the MCP server directly through the local development tool. The project supports a command pattern that runs the MCP server with a specified config.

uv run conductor-mcp --config <ABSOLUTE PATH TO A JSON CONFIG FILE>
```
> Note: a local_development.py configuration exists for setting environment variables when using the local development flow.