databricks-mcp

What it is and what it does

databricks-mcp is a lightweight integration library that bridges MCP (Model Context Protocol) servers with Databricks platforms. It provides OAuth authentication helpers that work across Databricks Notebooks, Model Serving endpoints, and local development environments using the Databricks CLI, reducing boilerplate for developers building AI applications that need to run in multiple Databricks contexts.

The package is designed for teams already using Databricks and MLflow who want to adopt MCP servers without reimplementing authentication and environment-specific setup. It sits between your MCP server code and Databricks' runtime, handling the credential and context plumbing so you can focus on the protocol logic itself.

Use it for:

• Authenticate MCP servers running in Databricks notebooks without manually managing Databricks CLI credentials.
• Deploy MCP servers to Databricks Model Serving with pre-configured OAuth and environment detection.
• Develop and test MCP servers locally while maintaining compatibility with Databricks authentication flows.
• Integrate MLflow-tracked models with MCP servers in a Databricks workspace.
• Build AI agents that use MCP for tool access while running on Databricks infrastructure.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides helpers and utilities to integrate MCP (Model Context Protocol) servers with Databricks environments, including OAuth authentication support across notebooks, model serving, and local development.

Yes, if you are building MCP servers for Databricks environments. The package solves a real integration gap—OAuth and environment detection across three distinct Databricks contexts—and carries no security vulnerabilities or license friction. Install friction is low and maintenance is active. Not relevant for non-Databricks MCP work.

Install

databricks-mcp on PyPI

pip install databricks-mcp

uv add databricks-mcp

poetry add databricks-mcp

Installing databricks-mcp

Before you install

Low install friction with a pure-Python wheel distribution. Active maintenance status with a recent release. Depends on databricks-sdk, mlflow, mcp, and databricks-ai-bridge, all of which are standard data/ML ecosystem packages.

License in practice

Licensed under Apache-2.0 (permissive), which allows commercial use, modification, and distribution with minimal restrictions—suitable for most production and proprietary projects.

Quickstart

pip install databricks-mcp

from databricks_mcp import OAuthProvider
# Use OAuth provider for authentication in Databricks environments

Requires Python 3.10 or later.

Verify before relying

• Whether the OAuth provider works identically across all three environments (notebooks, model serving, local) or has environment-specific limitations.
• What specific MCP server patterns or use cases are best supported by the included helpers.
• Whether databricks-ai-bridge is a required runtime or an optional integration point.

Package facts

Evidence: databricks_mcp-0.9.2-py3-none-any.whl

Tags

More Application Frameworks packages