Gumloop

The Gumloop MCP Server lets you interact with Gumloop’s automation platform through a standardized interface, enabling you to start, monitor, and manage automations, access workbooks and flows, work with files, and pass input parameters contextually for each run.

How to use

You will run automations by starting a saved flow, supply inputs for that flow, and then monitor the run status. You can list available saved flows and workbooks, fetch their input schemas, and upload or download files used in automations. When you need to execute a flow, you provide your user and project context so the automation runs with parameters tailored to your needs. Use a client that can call the MCP endpoints to perform these actions in sequence or on-demand.

Typical workflows include starting a saved automation with specific input values, checking the run status to track progress, and retrieving outputs or logs once the run finishes. You can also prepare files ahead of time by uploading them, then reference them during automation execution. If you want to run multiple files at once, you can upload several files in a single operation and download a zip when needed.

How to install

Prerequisites: you should have either the NPX method available through Node.js or Docker installed on your machine.

1. Ensure you have Node.js installed for NPX usage, or install Docker if you prefer running a container.

2. Obtain your Gumloop API key with access to the required features from your Gumloop workspace settings.

3. If you plan to run via NPX, prepare your config with your API key in the environment for the MCP client. If you plan to run via Docker, prepare the same API key to pass to the container as an environment variable.

4. Start using the MCP client to call the server’s tools (startAutomation, retrieveRunDetails, listSavedFlows, listWorkbooks, retrieveInputSchema, uploadFile, and others) as needed for your automation workflows.

Configuration and usage examples

{
  "mcpServers": {
    "gumloop": {
      "command": "npx",
      "args": [
        "-y",
        "gumloop-mcp-server"
      ],
      "env": {
        "GUMLOOP_API_KEY": "<YOUR_GUMLOOP_API_KEY>"
      }
    }
  }
}

{
  "mcpServers": {
    "gumloop": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "GUMLOOP_API_KEY",
        "gumloop-mcp-server"
      ],
      "env": {
        "GUMLOOP_API_KEY": "<YOUR_GUMLOOP_API_KEY>"
      }
    }
  }
}

Files and inputs

Upload files you want to use in automations with the appropriate file name and content, then reference those files in your input parameters. Retrieve input schemas for saved flows to understand required fields and types before starting a run.

To start an automation, supply the user_id, saved_item_id, and an optional list of pipeline_inputs that map input_name to value. You can then monitor the run with retrieveRunDetails and examine outputs and logs after completion.

Security and limits

Use a valid Gumloop API key with appropriate permissions. Be aware of Gumloop’s rate limits and usage quotas. File uploads are limited by what the API allows, and certain features may require specific subscription tiers.

Examples of common tasks

Starting an automation, checking its status, and working with files are common tasks you’ll perform frequently. You can upload “data.csv” via the file upload endpoint, then pass its reference as part of the inputs to a saved flow. You can later download outputs or any generated files tied to a run.

Available tools

startAutomation

Initiates a new flow run for a saved automation, providing user and flow identifiers and optional inputs to customize the run.

retrieveRunDetails

Fetches detailed information about a specific automation run, including state, outputs, timestamps, and logs.

listSavedFlows

Returns a list of all saved automation flows for a user or project, with metadata for each flow.

listWorkbooks

Provides a list of all workbooks and their associated saved flows for a user or project.

retrieveInputSchema

Retrieves the input schema for a specific saved flow, detailing required input parameters.

uploadFile

Uploads a single file to the Gumloop platform for use in automations.

uploadMultipleFiles

Uploads multiple files in a single request for use across automations.

downloadFile

Downloads a specific file associated with a flow run or saved item.

downloadMultipleFiles

Downloads multiple files as a zip archive for a given run.