Databricks MCP Server

Name: Databricks MCP Server
Author: RafaelCartenet

@RafaelCartenet

About Databricks MCP Server

Model Context Protocol (MCP) server for Databricks that empowers AI agents to autonomously interact with Unity Catalog metadata. Enables data discovery, lineage analysis, and intelligent SQL execution. Agents explore catalogs/schemas/tables, understand relationships, discover not

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "mcp-databricks-server-rafaelcartenet": {
      "command": "uv",
      "args": [
        "pip",
        "install",
        "-r",
        "requirements.txt"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Databricks MCP Server?

This server enables Large Language Models (LLMs) to interact with Databricks by providing tools for Unity Catalog (UC) metadata exploration, data lineage analysis, and SQL query execution. It is designed for developers and data engineers who want AI agents to autonomously discover, understand, and query Databricks data assets.

How to use Databricks MCP Server?

Install Python 3.10+ and dependencies (via pip install -r requirements.txt or uv). Set environment variables DATABRICKS_HOST, DATABRICKS_TOKEN, and DATABRICKS_SQL_WAREHOUSE_ID (via .env file or export). Run the server in standalone mode or integrate with Cursor using an MCP client configuration. The server exposes tools such as execute_sql_query, list_uc_catalogs, describe_uc_catalog, describe_uc_schema, and describe_uc_table.

Key features of Databricks MCP Server

Execute arbitrary SQL queries using Databricks SDK.
List and describe Unity Catalogs, schemas, and tables.
Retrieve comprehensive table lineage (tables, notebooks, jobs).
Read and analyze notebook content for transformation logic.
All descriptive tools return Markdown, optimized for LLM consumption.

Use cases of Databricks MCP Server

Data discovery: automatically explore catalogs and schemas to locate relevant tables.
Impact analysis: trace upstream/downstream dependencies of tables, notebooks, and jobs.
Code exploration: examine actual data transformation logic inside notebooks.
Autonomous SQL query generation: have an agent construct accurate queries using UC metadata.

FAQ from Databricks MCP Server

What is the motivation behind this server?

Documenting data assets in Unity Catalog requires effort; this server proves that investment pays off by letting LLM agents directly use that metadata to generate more intelligent and accurate SQL queries.

What permissions are required for the token?

The identity must have USE CATALOG and USE SCHEMA on the catalogs and schemas to be accessed, plus appropriate permissions for SQL warehouse operations when executing queries or fetching lineage.

How does the server handle long-running queries?

The execute_sql_query tool is designed for interactive queries; for long-running or resource-intensive queries, users should implement their own timeout or cancellation logic on the client side, as the server does not impose a built-in limit.

What are the dependencies?

Python 3.10+ and the packages listed in requirements.txt (or installed via uv). The Databricks SDK is used for UC APIs and SQL warehouse execution.

Where does data lineage come from?

Lineage is fetched from Unity Catalog and includes table-to-table relationships, notebooks that read from or write to tables, and associated Databricks job information. Notebook paths enable direct code reading within the workspace.

Comments

More Other MCP servers

Reactive Resume

amruthpillai

A one-of-a-kind resume builder that keeps your privacy in mind. Completely secure, customizable, portable, open-source and free forever. Try it out today!

39.5K