JordiNei/mcp-databricks-server

Overview

An open-source Python server that exposes Databricks workspace resources through the Model Context Protocol. It allows MCP-compatible AI agents to interact with Databricks clusters, jobs, and notebooks via a standardized interface.

Best for

Best for
Developers building agentic workflows that need read/write access to Databricks resources

Use cases

• Retrieve status of Databricks job runs from an agent
• List and inspect available clusters in a workspace
• Trigger notebook executions and check results programmatically

How to use

Install

pip install -r requirements.txt

Tools exposed

• run_sql_query
• list_jobs
• get_job_status
• get_job_details

Example client config

DATABRICKS_HOST=your-databricks-instance.cloud.databricks.com\nDATABRICKS_TOKEN=your-personal-access-token\nDATABRICKS_HTTP_PATH=/sql/1.0/warehouses/your-warehouse-id

Notes

An open-source Python server that exposes Databricks workspace resources through the Model Context Protocol. It allows MCP-compatible AI agents to interact with Databricks clusters, jobs, and notebooks via a standardized interface.

49 stars on GitHub. Last updated 2025-04-24.

Use cases

• Retrieve status of Databricks job runs from an agent
• List and inspect available clusters in a workspace
• Trigger notebook executions and check results programmatically

Pros

• Lightweight Python implementation easy to extend
• Compatible with any MCP client (e.g., Claude, custom agents)
• Open source with permissive licensing

Cons

• Small community (49 stars) and limited documentation
• Requires a Databricks API token and workspace URL
• Feature set likely basic; may miss advanced Databricks APIs

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.