Dataiku

You are using a dedicated MCP server that enables you to access and analyze Dataiku DSS REST APIs for flow analysis, project metadata, datasets, recipes, jobs, scenarios, folders, variables, connections, and code environments. It provides deterministic flow maps, clear status outputs, and structured error handling to support reliable day-to-day operations.

How to use

You can run your Dataiku MCP server from your development environment or via a client that speaks MCP. The recommended approach is to start the local MCP server using the provided npm-based workflow and then connect your MCP client to the server using the explicit environment configuration shown in the setup example.

To use the server from a client, configure the MCP client to invoke the server as an external process. The client should execute the following runtime configuration: use npx to launch the MCP server package with the required environment values for your DSS instance. The standard environment keys you provide are the DSS URL, your API key, and an optional default project key.

A practical workflow is to first install dependencies, build the server, and then run it locally. After startup, use the map action to obtain deterministic flow maps for your projects.

How to install

Prerequisites you need before installing the MCP server:

• Node.js 20 or newer

• npm

• Dataiku DSS URL and API key

Follow these steps to install and run locally from the repository root.

# install dependencies
npm ci

# static checks
npm run check

# unit tests
npm test

# build distribution
npm run build

# run MCP server locally (dev)
npm start

Environment and usage notes

To connect clients to the MCP server, you need to provide the Dataiku DSS base URL and an API key. You can optionally set a default project key to scope operations.

When you install the server, you typically create a local environment file and populate it with your credentials. A minimal setup uses the following environment keys.

Configuration references for clients

You can run the MCP server from an MCP client using the following configuration, which is designed to be used by command-line clients.

{
  "command": "npx",
  "args": ["-y", "dataiku-mcp"],
  "env": {
    "DATAIKU_URL": "https://your-dss-instance.app.dataiku.io",
    "DATAIKU_API_KEY": "your_api_key",
    "DATAIKU_PROJECT_KEY": "YOUR_PROJECT_KEY"
  }
}

Available tools

project

Operations to list, get, describe flow, and map projects to derive deterministic flow maps with recipe subtypes and connectivity.

dataset

Access dataset metadata, schema, previews, and data download, plus create, update, and delete datasets.

recipe

Manage recipes including list, get, create, update, delete, and download for recipe-level analysis.

job

Interact with jobs by listing, fetching logs, building, and aborting as part of workflow management.

scenario

Execute and monitor scenarios with list, run, status, get, create, update, and delete operations.

managed_folder

Handle folders and their contents, including upload and download of files within managed storage.

variable

Get and set variables to manage configuration and runtime behavior.

connection

Infer connections to data sources and systems used by Dataiku DSS flows.

code_env

List and retrieve code environments used to run recipes and notebooks.