JustTryAI_databricks-mcp-server

JustTryAI_databricks-mcp-server

0
cloud_platformsdeveloper_tools

The Databricks MCP Server is designed to enable interaction with Databricks clusters, jobs, and notebooks via the MCP protocol. It supports asynchronous operations and integrates seamlessly with Databricks' REST API, providing various tools for cluster and job management.

Databricks MCP Server

A Model Completion Protocol (MCP) server for Databricks that facilitates interaction with Databricks clusters, jobs, and notebooks through LLM-powered tools.

Features

  • MCP Protocol Support: Integrates the MCP protocol for LLM interaction.
  • Databricks API Integration: Access Databricks REST API functionalities.
  • Tool Registration: Offers Databricks operations as MCP tools.
  • Async Support: Utilizes asyncio for efficient operation.

Available Tools

  • list_clusters, create_cluster, terminate_cluster, get_cluster, start_cluster
  • list_jobs, run_job
  • list_notebooks, export_notebook
  • list_files, execute_sql

Installation

  • Prerequisites: Python 3.10+, uv package manager
  • Setup involves cloning the repo, installing dependencies, and setting environment variables.

Running the MCP Server

Execute start scripts available for Windows and Linux/Mac for server initiation and MCP protocol connection readiness.

Project Structure

The project includes directories for source code, tests, scripts, and documentation.

Development

Follows PEP 8 guidelines, uses specific linting tools, and employs pytest for testing with a goal of 80% coverage.

Documentation

Generated with Sphinx, with examples available in the examples/ directory.

Contributing

Contributions are welcome; adhere to coding standards and include tests and documentation.

License

Licensed under the MIT License.