Coupler.io

The Coupler.io MCP Server enables you to query and transform data from Coupler.io data flows, allowing you to analyze multi-channel marketing, sales, finance, and other business data within your preferred AI tooling. It connects to your Coupler.io data sources and exposes data so you can ask questions, fetch reports, and derive actionable insights from up-to-date information.

How to use

You can use the MCP server by configuring your MCP client to connect to the Coupler.io data flows. Once connected, you can query marketing, sales, and finance metrics as if you were collaborating with a data analyst. Ask for dashboards, forecasted revenue, pipeline reports, or channel ROI and receive structured results you can act on.

How to install

Prerequisites you need before starting are: install Docker and ensure it is running on your machine. You also need a Coupler.io Personal Access Token to authorize the MCP server.

{
  "mcpServers": {
    "coupler": {
      "command": "docker",
      "args": [
        "run",
        "--pull=always",
        "-e",
        "COUPLER_ACCESS_TOKEN",
        "--rm",
        "-i",
        "ghcr.io/railsware/coupler-io-mcp-server"
      ],
      "env": {
        "COUPLER_ACCESS_TOKEN": "<your_token>"
      }
    }
  }
}

Additional notes

You can also build and install a local MCP using a .mcpb file if you have a workflow that targets local development. A dedicated command sequence for creating and installing the .mcpb file is available in your workflow, and you can then load the MCP into your environment using your preferred client. Ensure you manage your token securely and revoke access if it is compromised.

Configuration and security

Manage access by keeping your Coupler.io Personal Access Token secure. When running in containers or IDEs, avoid exposing tokens in logs or public configuration files. If you rotate tokens, update the MCP client and restart the server to apply the new credentials.

Troubleshooting and debugging

If you encounter issues starting the server, verify Docker is running and that the environment variable COUPLER_ACCESS_TOKEN is set correctly. Check the container logs for errors related to authentication or data flow access. If the server is not reachable from your MCP client, confirm network access and that the token has the required permissions to read your data flows.

Available tools

get-data

Gets the result of a data flow run as a SQLite file and executes a read-only query on it, assuming the data flow has an AI destination.

get-schema

Fetches the data flow schema file for data flows built from a dashboard or dataset template.

list-dataflows

Lists data flows that have an AI destination.

get-dataflow

Retrieves metadata about a data flow, including sources, connections, last successful execution, and error details if present.