home / mcp / marketing connect mcp server

Marketing Connect MCP Server

Provides an MCP server that exposes tools, resources, and prompts for Marketing Connect AI integrations.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "alanedavis-mcp": {
      "command": "uv",
      "args": [
        "run",
        "marketing-connect-mcp",
        "--port",
        "8000"
      ],
      "env": {
        "MCP_HOST": "0.0.0.0",
        "MCP_PORT": "8000",
        "MCP_DEBUG": "false",
        "MCP_REGION": "us-east-1",
        "MCP_BASE_URL": "https://your-app-url.com",
        "MCP_LOG_LEVEL": "INFO",
        "MCP_SERVER_NAME": "marketing-connect-mcp-services",
        "MCP_SERVER_VERSION": "1.0.0"
      }
    }
  }
}

You run Marketing Connect MCP Services to enable AI models to securely interact with external tools, data, and templates. This MCP server provides a framework for exposing tools, resources, and prompts that your AI clients can invoke within a standardized protocol.

How to use

You use the MCP server by running it locally or deploying it to your environment, then connecting with an MCP client. The server exposes a set of endpoints and transport options that let the AI request tool invocations, load contextual data, and utilize reusable prompts. Start a run in your environment, then initialize an MCP session from your client, making subsequent requests to list tools, call tools, or query server information as needed.

As you develop or integrate, your MCP client will establish a session, learn what tools are available, and invoke specific tools with the appropriate inputs. The server provides health checks and a simple info endpoint so you can verify deployment status and version information at a glance.

How to install

Prerequisites you need before installation include a Python 3.11+ environment and a FastMCP-compatible runtime. You also need a shell with standard tooling to run build and start commands.

Step by step, install and run the server as follows. Copy these commands into your terminal in sequence.

make ci-prebuild
make build
uv run marketing-connect-mcp --port 8000
# Alternatively start with the default in a typical deployment
make run
# Or with debug logging
make run-debug

Configuration and notes

Configuration is controlled via environment variables prefixed with MCP_. Copy the example template and customize values for your deployment. Common settings include server identity, port, and logging. The following variables are typical: server name, server version, host, port, debug mode, log level, base URL, and region.

If you are running in a container or cloud environment, set the following example values to reflect your deployment: MCP_SERVER_NAME, MCP_SERVER_VERSION, MCP_HOST, MCP_PORT, MCP_DEBUG, MCP_LOG_LEVEL, MCP_BASE_URL, MCP_REGION. Adjust these to match your operational needs.

Health and status endpoints are available to confirm your deployment is up and healthy. Use the health check to verify readiness, and query the info endpoint to retrieve version and uptime details.

Troubleshooting and notes

If you encounter issues, ensure the MCP server process is running and reachable at the configured port. Check the health endpoint for a quick status, and review log output if debug mode is enabled. Ensure your MCP client is configured to point to the correct host and port.

For local development, you can run the server in debug mode to get more verbose logs, which helps diagnose issues during integration with your AI clients.