Databricks

关于

This is a robust and fixed version of the Databricks Model Completion Protocol (MCP) server, designed to seamlessly integrate Databricks functionalities with LLM-powered tools like Claude Code. It resolves critical issues found in the original implementation, such as asyncio conflicts, command spawning limitations, and SQL API deficiencies, ensuring reliable and efficient operation. The server exposes comprehensive Databricks features, including cluster, job, notebook, file system, and SQL execution management, empowering LLMs to automate complex data workflows and infrastructure interactions.

主要功能

• Comprehensive Databricks API Integration: Accesses 20 tools across cluster, job, notebook, file system, and SQL management.
• Seamless LLM Integration: Implements the Model Completion Protocol for direct interaction with LLM clients like Claude Code.
• Serverless Compute Support: Create jobs with serverless execution by default, simplifying resource management and optimizing costs.
• Advanced Job & SQL Management: Includes pagination, filtering, run status tracking for jobs, and non-blocking SQL execution with status monitoring.
• Unity Catalog & DBFS File Operations: Supports robust file uploads and listings for both Unity Catalog volumes and DBFS.
• 0 GitHub stars

使用案例

• Automate Databricks operations (e.g., creating clusters, running jobs, managing notebooks) using natural language commands from LLMs.
• Execute and monitor complex, long-running SQL queries asynchronously within a Databricks environment.
• Streamline data ingestion and management by uploading files directly to Unity Catalog volumes or DBFS through an LLM interface.