Databricks MCP Server

Official

by pramodbhatofficial · Python

263 tools for Databricks: Unity Catalog, SQL, Compute, Jobs, Serving, and more.

Databricks MCP server that exposes a model context protocol for Databricks tooling. It aggregates 263 tools spanning Unity Catalog, SQL, Compute, Jobs, Serving, MLflow, and more, enabling integration with AI workflows and LLM-assisted operations.

🛠️ Key Features

263 tools exposed for Databricks platforms
Covers Unity Catalog, SQL, Compute, Jobs, Serving, and ML workflows
Model Context Protocol (MCP) compatibility for context-aware tooling
Python-based server with Apache 2.0 license
Readme excerpt provided for quick onboarding

🚀 Use Cases

Contextual model interactions within Databricks lakehouse
Tooling for MLflow and ML lifecycle automation
Integration for AI assistants with Databricks-native resources
Embedding Databricks capabilities into LLM-powered workflows

⚡ Developer Benefits

Rich tool surface for MCP-backed orchestration
Clear README excerpts and repository signals for quick setup
Open-source style visibility with CI and PyPI badges

⚠️ Limitations

Source data excerpt is partial; verify full feature set from repository
Compatibility asserted via MCP standards; ensure environment supports all tools
License and dependency details require direct repo review

Topics

aiclaudedatabricksdatabricks-appslakebasellmmcpmlflowmodel-context-protocolpythonsdksqlunity-catalogvector-searchlakehousemlops

README.md12 KBView on GitHub

Databricks MCP Server

A comprehensive Model Context Protocol (MCP) server for Databricks, built on the official Databricks Python SDK.

Provides 263 tools and 8 prompt templates across 28 service domains, giving AI assistants full access to the Databricks platform.

Features

SDK-first: Uses databricks-sdk for type safety and automatic API freshness
Comprehensive: Covers Unity Catalog, SQL, Compute, Jobs, Pipelines, Serving, Vector Search, Apps, Lakebase, Dashboards, Genie, Secrets, IAM, Connections, Experiments, and Delta Sharing
Zero custom auth: Delegates authentication entirely to the SDK (PAT, OAuth, Azure AD, service principal -- all automatic)
Selective loading: Include/exclude tool modules via environment variables
MCP Resources: Read-only workspace context (URL, current user, auth type)

Quick Start

Installation

bash

pip install databricks-sdk-mcp

Or run with Docker:

bash

docker run -i -e DATABRICKS_HOST=... -e DATABRICKS_TOKEN=... databricks-mcp

Or install from source:

bash

git clone https://github.com/pramodbhatofficial/databricks-mcp-server.git
cd databricks-mcp-server
pip install -e ".[dev]"

Authentication

Authentication is handled by the Databricks SDK. Set one of:

Personal Access Token (simplest):

bash

export DATABRICKS_HOST=https://your-workspace.databricks.com
export DATABRICKS_TOKEN=dapi...

OAuth (M2M):

bash

export DATABRICKS_HOST=https://your-workspace.databricks.com
export DATABRICKS_CLIENT_ID=...
export DATABRICKS_CLIENT_SECRET=...

Other methods: Azure AD, Databricks CLI profile, Azure Managed Identity -- all auto-detected by the SDK.

Running

bash

databricks-mcp

This starts the MCP server using stdio transport.

Integrations

Claude Code (Terminal)

Add to ~/.claude/settings.json or your project's .claude/settings.json:

json

{
  "mcpServers": {
    "databricks": {
      "command": "databricks-mcp",
      "env": {
        "DATABRICKS_HOST": "https://your-workspace.databricks.com",
        "DATABRICKS_TOKEN": "dapi..."
      }
    }
  }
}

Then restart Claude Code. Verify with /mcp to see the registered tools.

Claude Desktop

Add to your Claude Desktop config file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

json

{
  "mcpServers": {
    "databricks": {
      "command": "databricks-mcp",
      "env": {
        "DATABRICKS_HOST": "https://your-workspace.databricks.com",
        "DATABRICKS_TOKEN": "dapi..."
      }
    }
  }
}

Restart Claude Desktop. The Databricks tools will appear in the tool picker.

Cursor

Add to .cursor/mcp.json in your project root (or ~/.cursor/mcp.json for global):

json

{
  "mcpServers": {
    "databricks": {
      "command": "databricks-mcp",
      "env": {
        "DATABRICKS_HOST": "https://your-workspace.databricks.com",
        "DATABRICKS_TOKEN": "dapi..."
      }
    }
  }
}

Open Cursor Settings > MCP to verify the server is connected.

Windsurf

Add to ~/.codeium/windsurf/mcp_config.json:

json

{
  "mcpServers": {
    "databricks": {
      "command": "databricks-mcp",
      "env": {
        "DATABRICKS_HOST": "https://your-workspace.databricks.com",
        "DATABRICKS_TOKEN": "dapi..."
      }
    }
  }
}

VS Code (Copilot)

Add to .vscode/mcp.json in your project:

json

{
  "servers": {
    "databricks": {
      "command": "databricks-mcp",
      "env": {
        "DATABRICKS_HOST": "https://your-workspace.databricks.com",
        "DATABRICKS_TOKEN": "dapi..."
      }
    }
  }
}

Zed

Add to Zed's settings (~/.config/zed/settings.json):

json

{
  "context_servers": {
    "databricks": {
      "command": {
        "path": "databricks-mcp",
        "env": {
          "DATABRICKS_HOST": "https://your-workspace.databricks.com",
          "DATABRICKS_TOKEN": "dapi..."
        }
      }
    }
  }
}

Any MCP Client (Generic stdio)

The server uses stdio transport. Connect from any MCP-compatible client:

bash

# Set auth env vars
export DATABRICKS_HOST=https://your-workspace.databricks.com
export DATABRICKS_TOKEN=dapi...

# Start the server (communicates via stdin/stdout)
databricks-mcp

Tip: Load Only What You Need

If your MCP client struggles with many tools, use selective loading to reduce the tool count:

json

{
  "mcpServers": {
    "databricks": {
      "command": "databricks-mcp",
      "env": {
        "DATABRICKS_HOST": "https://your-workspace.databricks.com",
        "DATABRICKS_TOKEN": "dapi...",
        "DATABRICKS_MCP_TOOLS_INCLUDE": "unity_catalog,sql,compute,jobs"
      }
    }
  }
}

Tool Modules

Module	Tools	Description
`unity_catalog`	23	Catalogs, schemas, tables, volumes, functions, registered models
`sql`	14	Warehouses, SQL execution, queries, alerts, history
`workspace`	10	Notebooks, files, repos
`compute`	18	Clusters, instance pools, policies, node types, Spark versions
`jobs`	13	Jobs, runs, tasks, repair, cancel all
`pipelines`	8	DLT / Lakeflow pipelines
`serving`	10	Serving endpoints, model versions, OpenAPI
`vector_search`	10	Vector search endpoints, indexes, sync
`apps`	10	Databricks Apps lifecycle
`database`	10	Lakebase PostgreSQL instances
`dashboards`	9	Lakeview AI/BI dashboards, published views
`genie`	5	Genie AI/BI conversations
`secrets`	8	Secret scopes and secrets
`iam`	16	Users, groups, service principals, permissions, current user
`connections`	5	External connections
`experiments`	14	MLflow experiments, runs, artifacts, metrics, params
`sharing`	11	Delta Sharing shares, recipients, providers
`files`	12	DBFS and UC Volumes file operations
`grants`	3	Unity Catalog permission grants (GRANT/REVOKE)
`storage`	10	Storage credentials and external locations
`metastores`	8	Unity Catalog metastore management
`online_tables`	3	Online tables for low-latency serving
`global_init_scripts`	5	Workspace-wide init scripts
`tokens`	5	Personal access token management
`git_credentials`	5	Git credential management for repos
`quality_monitors`	8	Data quality monitoring and refreshes
`command_execution`	4	Interactive command execution on clusters
`workflows`	5	Composite multi-step operations (workspace status, schema setup, query preview)

Selective Tool Loading

With 263 tools, it's recommended to load only the modules you need. This improves agent performance and tool selection accuracy.

Role-Based Presets (Recommended)

Pick a preset that matches your role:

Preset	Modules	Tools	Config
Data Engineer	unity_catalog, sql, compute, jobs, pipelines, files, quality_monitors	~100	`DATABRICKS_MCP_TOOLS_INCLUDE=unity_catalog,sql,compute,jobs,pipelines,files,quality_monitors`
ML Engineer	serving, vector_search, experiments, compute, unity_catalog, online_tables, files	~98	`DATABRICKS_MCP_TOOLS_INCLUDE=serving,vector_search,experiments,compute,unity_catalog,online_tables,files`
Platform Admin	iam, secrets, tokens, metastores, compute, global_init_scripts, grants, storage	~85	`DATABRICKS_MCP_TOOLS_INCLUDE=iam,secrets,tokens,metastores,compute,global_init_scripts,grants,storage`
App Developer	apps, database, sql, files, serving, secrets	~64	`DATABRICKS_MCP_TOOLS_INCLUDE=apps,database,sql,files,serving,secrets`
Data Analyst	sql, unity_catalog, dashboards, genie, workspace	~61	`DATABRICKS_MCP_TOOLS_INCLUDE=sql,unity_catalog,dashboards,genie,workspace`
Minimal	sql, unity_catalog	~37	`DATABRICKS_MCP_TOOLS_INCLUDE=sql,unity_catalog`

Example using a preset in Claude Code:

json

{
  "mcpServers": {
    "databricks": {
      "command": "databricks-mcp",
      "env": {
        "DATABRICKS_HOST": "https://your-workspace.databricks.com",
        "DATABRICKS_TOKEN": "dapi...",
        "DATABRICKS_MCP_TOOLS_INCLUDE": "unity_catalog,sql,compute,jobs,pipelines,files,quality_monitors"
      }
    }
  }
}

Custom Filtering

bash

# Only include specific modules
export DATABRICKS_MCP_TOOLS_INCLUDE=unity_catalog,sql,serving

# Exclude specific modules (cannot combine with INCLUDE)
export DATABRICKS_MCP_TOOLS_EXCLUDE=iam,sharing,experiments

If INCLUDE is set, only those modules load. If EXCLUDE is set, everything except those modules loads. INCLUDE takes precedence if both are set.

Tool Discovery (For AI Agents)

The server includes built-in tool discovery to help AI agents find the right tools:

MCP Resources

URI	Description
`databricks://workspace/info`	Workspace URL, current user, auth type
`databricks://tools/guide`	Tool catalog with module descriptions, use cases, and role presets

Agents can read databricks://tools/guide at connection time to understand what's available.

Discovery Tool

The databricks_tool_guide tool helps agents find the right tools during a conversation:

code

# Find tools for a specific task
databricks_tool_guide(task="run a SQL query")
databricks_tool_guide(task="deploy an ML model")
databricks_tool_guide(task="create a user")

# Get role-based recommendations
databricks_tool_guide(role="data_engineer")
databricks_tool_guide(role="ml_engineer")

This returns matching modules with descriptions and usage hints, so the agent knows exactly which databricks_* tools to call.

MCP Prompts (Guided Workflows)

The server includes 8 prompt templates that guide AI agents through multi-step Databricks workflows:

Prompt	Description
`explore_data_catalog`	Browse Unity Catalog structure (catalogs → schemas → tables)
`query_data`	Find a warehouse, execute SQL, and format results
`debug_failing_job`	Investigate a failing job: status, logs, error analysis
`setup_ml_experiment`	Create an MLflow experiment and configure tracking
`deploy_model`	Deploy a model to a serving endpoint
`setup_data_pipeline`	Create a DLT pipeline with scheduling
`workspace_health_check`	Audit clusters, warehouses, jobs, and endpoints
`manage_permissions`	Review and update permissions on workspace objects

Prompts appear automatically in MCP clients that support them (e.g., Claude Desktop's prompt picker).

Docker

Run the MCP server in a container:

bash

# Build
docker build -t databricks-mcp .

# Run with stdio
docker run -i \
  -e DATABRICKS_HOST=https://your-workspace.databricks.com \
  -e DATABRICKS_TOKEN=dapi... \
  databricks-mcp

# Run with SSE transport
docker run -p 8080:8080 \
  -e DATABRICKS_HOST=https://your-workspace.databricks.com \
  -e DATABRICKS_TOKEN=dapi... \
  databricks-mcp --transport sse --port 8080

# Run with selective modules
docker run -i \
  -e DATABRICKS_HOST=https://your-workspace.databricks.com \
  -e DATABRICKS_TOKEN=dapi... \
  -e DATABRICKS_MCP_TOOLS_INCLUDE=sql,unity_catalog \
  databricks-mcp

SSE Transport (Remote Server)

The server supports SSE transport for remote connections:

bash

# Start as SSE server
databricks-mcp --transport sse --port 8080

# Custom host/port
databricks-mcp --transport sse --host 127.0.0.1 --port 3000

Connect from any MCP client that supports SSE:

json

{
  "mcpServers": {
    "databricks": {
      "url": "http://localhost:8080/sse"
    }
  }
}

Development

bash

# Install with dev dependencies
pip install -e ".[dev]"

# Lint
ruff check databricks_mcp/

# Test
pytest tests/ -v

Author

Pramod Bhat

License

Apache 2.0 -- see LICENSE.

Install

Configuration

Environment variables

DATABRICKS_HOSTrequired

Your Databricks workspace URL (e.g. https://adb-123.azuredatabricks.net)

DATABRICKS_TOKENrequiredsecret

Databricks personal access token or OAuth token

DATABRICKS_MCP_TOOLS_INCLUDE

Comma-separated list of tool modules to load (e.g. unity_catalog,sql,compute). Leave empty to load all 263 tools.

Add to VS Code Add to Cursor

claude_desktop_config.json

json

{
  "mcpServers": {
    "databricks-sdk-mcp": {
      "command": "uvx",
      "args": [
        "databricks-sdk-mcp"
      ],
      "env": {
        "DATABRICKS_HOST": "<YOUR_DATABRICKS_HOST>",
        "DATABRICKS_TOKEN": "<YOUR_DATABRICKS_TOKEN>",
        "DATABRICKS_MCP_TOOLS_INCLUDE": "<YOUR_DATABRICKS_MCP_TOOLS_INCLUDE>"
      }
    }
  }
}

🛠️ Key Features

🚀 Use Cases

⚡ Developer Benefits

⚠️ Limitations

Topics

Databricks MCP Server

Features

Quick Start

Installation

Authentication

Running

Integrations

Claude Code (Terminal)

Claude Desktop

Cursor

Windsurf

VS Code (Copilot)

Zed

Any MCP Client (Generic stdio)

Tip: Load Only What You Need

Tool Modules

Selective Tool Loading

Role-Based Presets (Recommended)

Custom Filtering

Tool Discovery (For AI Agents)

MCP Resources

Discovery Tool

MCP Prompts (Guided Workflows)

Docker

SSE Transport (Remote Server)

Development

Author

License

Install

Configuration

Related servers