Databricks Agent Tools
Connect AI agents to Databricks for enterprise data access and developer automation.
About
This tool provides an experimental collection of Model Context Protocol (MCP) servers designed to enhance AI agents' interaction within the Databricks ecosystem. It enables agents to fetch enterprise data from Unity Catalog by exposing vector search indexes, functions, and Genie spaces as actionable tools. While the Unity Catalog server implementation in this repository is considered deprecated in favor of Databricks Managed MCP servers, it demonstrates how to build and deploy custom MCP servers on Databricks Apps for data access and to develop tools for automating common developer actions on the platform.
Key Features
- Supports interaction with Databricks Genie spaces for conversational AI
- Exposes Unity Catalog functions and vector search indexes as agent tools
- Enables deployment as a Databricks App for scalable hosting
- 76 GitHub stars
- Provides local usage via stdio transport for development and testing
- Integrates AI agents with Databricks Unity Catalog for data access
Use Cases
- Allow AI agents to query and retrieve structured and unstructured data from Unity Catalog
- Enable AI agents to interact with Databricks Genie spaces for conversational AI
- Automate common developer actions on Databricks using AI agents
Sponsored