Coupler.io MCP

How to Install and Use Coupler.io MCP

Coupler.io MCP is a Model Context Protocol (MCP) server designed to help you connect and analyze business data quickly and easily. It works with Coupler.io data flows, letting you query marketing, sales, and finance data from many sources. Below is a simple guide to get you started with installation and running the Coupler.io MCP.

Preparing to Install Coupler.io MCP

Before you start, you need to have a few things ready. These are important to make sure the server runs properly.

First, install Docker on your computer if you haven't yet. Docker helps you run the Coupler.io MCP server in a container safely and easily. You can download Docker from docker.com.

Next, make sure Docker is running on your machine.

Finally, get your Coupler.io Personal Access Token. This token is needed to connect your Coupler.io data flows with the server. You can find or create this token in your Coupler.io account at their MCP app page.

Once you have these, you are ready to run the server.

Running Coupler.io MCP Using Docker

To run the Coupler.io MCP server, you will use Docker commands. They make it simple to start the server with your access token securely set.

Here is an example of a Docker command to run the server:

docker run --pull=always -e COUPLER_ACCESS_TOKEN=<your_token> --rm -i ghcr.io/railsware/coupler-io-mcp-server

• Replace <your_token> with your actual Coupler.io Personal Access Token.
• The --pull=always option makes sure you use the newest version of the server image every time you start it.
• --rm removes the container when you stop it.
• -i keeps the container open and interactive.

This command starts the server, and then you can start asking questions or querying your Coupler.io data flows through AI tools like Claude.

Using Coupler.io MCP with Claude Desktop

If you use Claude Desktop, you can add Coupler.io MCP as a data server for easy access. You just need to add a config entry like this in your claude_desktop_config.json file:

{
  "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>"
      }
    }
  }
}

• Again, replace <your_token> with your Coupler.io access token.
• This setup runs the Docker container automatically inside Claude Desktop.

Exploring Tools in Coupler.io MCP

With the server running, you have tools available to work with your data flows. Here’s what you can do:

• Get Data: Run a query to get results from a data flow run on SQLite files.
• Get Schema: Fetch the schema (structure) of a specific data flow.
• List Dataflows: See all your data flows that have AI destinations.
• Get Dataflow: View metadata about a data flow, including its sources and status.

When using these tools, you will provide required inputs like dataflowId and executionId, which you get from your Coupler.io environment.

Summary

1. Install and run Docker.
2. Get your Coupler.io Personal Access Token.
3. Run the MCP server with Docker, setting your token as shown.
4. Optionally, configure Claude Desktop using the provided JSON.
5. Use the built-in tools to query and analyze your Coupler.io data.

This simple setup allows you to connect your business data flows to AI tools easily. Coupler.io MCP helps make your data smarter and your decisions faster.