An MCP server that enables AI agents to interact with Databricks Unity Catalog metadata, execute SQL queries, and explore data lineage, including notebook and task dependencies for comprehensive data analysis.

Installation

Pip

pip install -r requirements.txt

UV

uv pip install -r requirements.txt

Configuration

Cursor (UV)

{
    "mcpServers": {
        "databricks": {
            "command": "uv",
            "args": [
                "--directory",
                "/path/to/your/mcp-databricks-server",
                "run",
                "main.py"
            ]
        }
    }
}

Cursor (Python)

{
    "mcpServers": {
        "databricks": {
            "command": "python",
            "args": [
                "/path/to/your/mcp-databricks-server/main.py"
            ]
        }
    }
}

Available Tools

Tool	Description
list_uc_catalogs	Lists all available Unity Catalogs with their names, descriptions, and types
describe_uc_catalog	Provides an overview of a specific Unity Catalog, listing all its schemas with names and descriptions...
describe_uc_schema	Provides detailed information about a specific schema in Unity Catalog, optionally including...
describe_uc_table	Provides a detailed description of a Unity Catalog table with comprehensive lineage capabilities in...
execute_sql_query	Executes SQL queries against Databricks SQL warehouse and returns formatted results

Capabilities

• Execute arbitrary SQL queries using the Databricks SDK
• Markdown output optimized for LLMs for all descriptive tools
• Comprehensive Unity Catalog exploration and metadata access
• Data lineage analysis, including relationships between tables
• Discovery of notebook and task dependencies
• Code-level understanding through notebook content exploration
• End-to-end data flow analysis—from ingestion to consumption
• Impact analysis and debugging of data pipeline issues

Environment Variables

Required

• DATABRICKS_HOST - Hostname of your Databricks instance
• DATABRICKS_TOKEN - Your personal Databricks access token
• DATABRICKS_SQL_WAREHOUSE_ID - SQL Warehouse ID for executing queries and retrieving lineage

Usage Examples

Find and explore available data catalogs and schemas

Understand table structures and column details before querying

Track data lineage to understand upstream and downstream dependencies

Discover notebooks and tasks that process specific tables

Execute complex SQL queries for analysis and data retrieval

Notes

Requires Python 3.10+. The identity associated with DATABRICKS_TOKEN must have appropriate Unity Catalog permissions (USE CATALOG, USE SCHEMA, SELECT on tables) and CAN_USE permission on the SQL Warehouse. For production use, a service principal with narrowly scoped permissions is recommended.