databricks-mcp

🚀 MCP Databricks

A powerful Databricks integration for AI assistants via Model Context Protocol

📖 Introduction

MCP Databricks seamlessly connects AI assistants to your Databricks workspaces through the Model Context Protocol (MCP). Built with Python, it provides a rich collection of tools for managing virtually every aspect of your Databricks environment.

With this server, AI assistants like Claude can:

• 🔧 Manage compute resources with precision
• 📊 Execute SQL queries and analyze results
• 📁 Organize and manipulate workspace objects
• ✨ And much more!

🔍 Prerequisites

• 🐍 Python 3.11 or higher
• 💻 A Databricks workspace
• 🔑 Databricks Personal Access Token (PAT)
• 📦 Required Python packages (installed in setup)

🚀 Quickstart

1️⃣ Clone the repository

git clone https://github.com/leminkhoa/databricks-mcp
cd databricks-mcp

2️⃣ Configure environment variables

Create a .env file in the project root with your Databricks credentials:

# Databricks API Configuration
DATABRICKS_HOST="https://adb-<your workspace uri>.azuredatabricks.net/"
DATABRICKS_TOKEN="dapi_<your_token_here>"

# Server Configuration
SERVER_HOST="0.0.0.0"
SERVER_PORT="8000"
DEBUG="false"
TRANSPORT="stdio"

# Logging Configuration
LOG_LEVEL="INFO"

💡 Tip: For further instructions, You can use the env.sample file as a template.

3️⃣ Choose Your Installation Method

You can use MCP Databricks in two ways:

Option A: Docker (Recommended for production)

1. Build the Docker image:

docker build -t databricks-mcp .

2. Configure in Cursor with the following mcp.json entry:

{
  "databricks-mcp-docker": {
    "command": "docker",
    "args": [
      "run",
      "--rm",
      "-i",
      "--name", "databricks-mcp",
      "--env-file", "<path/to/.env>",
      "databricks-mcp"
    ]
  }
}

💡 Note: Replace <path/to/.env> with the absolute path to your .env file.

Option B: Local Installation with uv

1. Install uv (if not already installed):

curl -LsSf https://astral.sh/uv/install.sh | sh

💡 Note: See uv installation documentation for alternative installation methods.

2. Set up virtual environment:

uv venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install dependencies with uv:

uv sync

4. Configure in Cursor with the following mcp.json entry:

{
  "databricks-mcp-stdio": {
    "command": "uv",
    "args": [
      "--directory",
      "<repository directory>",
      "run",
      "main.py"
    ]
  }
}

💡 Note: Replace <repository directory> with the absolute path to your cloned repository.

🚀 Usage

Running the MCP Server

If not using Docker or Cursor integration, start the server with:

python main.py

or

uv run main.py

Connecting to Claude or other MCP clients

This server uses the stdio transport for seamless compatibility with Claude Desktop and other MCP clients. After installing the server, you can immediately connect to it using your preferred MCP client.

🧰 Tools and Capabilities

The MCP Databricks server provides a comprehensive toolkit for managing your Databricks environment:

💻 Cluster Management

📦 Library Management

🖥️ Command Execution

📊 SQL Warehouse Management

📁 Workspace Objects

---