🚀 MCP Databricks

What is 🚀 MCP Databricks?

🚀 MCP Databricks is a Python-based MCP server that connects AI assistants (like Claude) to Databricks workspaces. It provides a rich collection of tools for managing compute resources, executing SQL queries, and organizing workspace objects via the Model Context Protocol.

How to use 🚀 MCP Databricks?

Clone the repository, create a .env file with your DATABRICKS_HOST and DATABRICKS_TOKEN, then run the server using Docker (recommended for production) or locally with uv. Configure it in Cursor via mcp.json or connect directly to any MCP client (e.g., Claude Desktop) using stdio transport.

Key features of 🚀 MCP Databricks

• Manage Databricks clusters (create, start, delete, list)
• Execute Python, Scala, and SQL commands on clusters
• Install libraries (JAR, WHL, PyPI, Maven, CRAN)
• Manage SQL warehouses (list, create)
• Manipulate workspace objects (import, delete, create directories)

Use cases of 🚀 MCP Databricks

• Automate cluster lifecycle management via AI assistant conversation
• Run ad‑hoc SQL queries and analyze results in Databricks
• Install required libraries on running clusters interactively
• Organize and import notebooks or files into the workspace

FAQ from 🚀 MCP Databricks

What are the prerequisites for using 🚀 MCP Databricks?

You need Python 3.11 or higher, a Databricks workspace, and a Databricks Personal Access Token (PAT).

How do I install 🚀 MCP Databricks?

You can either build the Docker image with docker build -t databricks-mcp . or install locally with uv venv, uv sync, and then run uv run main.py.

How do I configure credentials?

Create a .env file in the project root with DATABRICKS_HOST (your workspace URL) and DATABRICKS_TOKEN (your PAT). Optional settings include server host, port, debug mode, and log level.

What transport does 🚀 MCP Databricks use?

The server uses the stdio transport for seamless compatibility with Claude Desktop and other MCP clients.

What tools does 🚀 MCP Databricks provide?

It offers tools for cluster management (list_clusters, create_cluster, delete_cluster, start_cluster, list_node_types, list_spark_versions, get_cluster), library installation (install_libraries), command execution (execute_command, create_execution_context), SQL warehouse management (list_sql_warehouses, create_sql_warehouse), and workspace object management (delete_workspace_object, get_workspace_object_status, import_workspace_object, create_workspace_directory).