Registry / devops / databricks-mcp

databricks-mcp

JSON →
library0.9.2pypypi✓ verified 27d ago

The `databricks-mcp` library provides helper utilities specifically designed for developers interacting with Databricks Workflows and the Databricks Control Plane (MCP). It wraps and extends the `databricks-sdk` to offer higher-level abstractions for common MCP operations, such as deployment stack management. Currently at version `0.9.0`, it is actively developed by DatabricksLabs, with releases tied to internal needs and `databricks-sdk` updates, typically on an infrequent basis.

pip install databricks-mcp
INSTALL
IMPORT
SIG · DATABRICKS-MCP
D
databricks-mcp
devopspythonv0.9.2
Install
40.5s avg
Import
12260ms
Disk
814MB
Pass rate
4/ 10
Env Coverage4 / 10
glibc
3.9–3.13
musl
3.9–3.13
Install & Compatibility
Where this runs
tested against v0.9.2 · pip install
no network on importno background threads
Install × environment matrix
Each cell = how many times install + import succeeded across repeated harness runs. Partial = flaky.
glibc = Debian/Ubuntu slim · musl = Alpine Linux
musl
glibc
py 3.10
✕ build_error
✓ 42.8s
py 3.11
✕ build_error
✓ 41.9s
py 3.12
✕ build_error
✓ 38.4s
py 3.13
✕ build_error
✓ 39s
py 3.9
✕ build_error
✕ build_error
814MB installed
● package 814MB
Code
Verified usage

Verified import paths — ran on the pinned version, not inferred.

DatabricksMCPClient
✓ from databricks_mcp import DatabricksMCPClient
✗ from databricks_mcp.databricks_mcp_client import DatabricksMcpClient
DatabricksOAuthClientProvider
✓ from databricks_mcp import DatabricksOAuthClientProvider
register_mcp_server_via_dcr
✓ from databricks_mcp import register_mcp_server_via_dcr

Initialize the `DatabricksMcpClient` by providing configuration, typically sourced from environment variables. This example demonstrates basic client instantiation, which validates your host and token against expected formats but doesn't make an external API call by default unless you interact with `client.account_client` or similar.

import os from databricks_mcp.databricks_config import DatabricksMcpConfig from databricks_mcp.databricks_mcp_client import DatabricksMcpClient # Ensure DATABRICKS_HOST, DATABRICKS_TOKEN, and DATABRICKS_ACCOUNT_ID # are set as environment variables for a real connection. # Token should be a Databricks Account-level token. try: config = DatabricksMcpConfig( host=os.environ.get("DATABRICKS_HOST", "https://accounts.cloud.databricks.com"), token=os.environ.get("DATABRICKS_TOKEN", "dapi-example-token"), account_id=os.environ.get("DATABRICKS_ACCOUNT_ID", "00000000-0000-0000-0000-000000000000"), ) client = DatabricksMcpClient(config) print(f"Databricks MCP Client initialized successfully for host: {client.config.host}") # The client object contains underlying databricks_sdk clients (e.g., client.account_client) # which can be used to perform actual operations. This quickstart only demonstrates initialization. except Exception as e: print(f"Failed to initialize Databricks MCP Client: {e}") print("Please ensure DATABRICKS_HOST, DATABRICKS_TOKEN, and DATABRICKS_ACCOUNT_ID") print("environment variables are correctly configured and point to your Databricks Account.")
databricks --version
Debug
Known issues
breakingThe library is currently at version `0.x.x` (pre-1.0), meaning API stability is not guaranteed. Breaking changes may be introduced in minor or patch releases without strict semantic versioning.
fix
Always pin exact versions in production environments (`databricks-mcp==0.9.0`) and thoroughly test updates before deployment.
affects: <1.0.0
gotchaAuthentication requires specific environment variables (`DATABRICKS_HOST`, `DATABRICKS_TOKEN`, `DATABRICKS_ACCOUNT_ID`). A common mistake is using a Databricks Workspace token instead of an Account-level token, which is necessary for Control Plane operations.
fix
Ensure `DATABRICKS_TOKEN` is an Account-level token obtained from your Databricks Account settings and that all three environment variables are correctly set before initializing `DatabricksMcpConfig`.
affects: All
gotcha`databricks-mcp` is a wrapper around `databricks-sdk`. If `databricks-sdk` introduces breaking changes or significant behavioral shifts that `databricks-mcp` does not immediately accommodate, it can lead to unexpected errors.
fix
Keep `databricks-mcp` and `databricks-sdk` versions synchronized and regularly check the release notes for both libraries. Consider pinning specific versions of `databricks-sdk` if stability is critical.
affects: All
gotchaThis library is primarily for Databricks Control Plane (MCP) interactions and complex workflows. For simpler workspace-level operations (e.g., managing notebooks, jobs, clusters within a single workspace), directly using the `databricks-sdk` might be more appropriate and less complex.
fix
Evaluate your use case; if it doesn't involve account-level management or complex deployment stacks across workspaces/accounts, the `databricks-sdk` might suffice.
affects: All
Upgrade
Version history
0.9.2latest on PyPI · released Jul 31, 2026
Audit
Dependencies
pythonrequiredRequired Python version
databricks-sdkrequiredCore dependency for interacting with Databricks APIs
Agent activity
25 hits · last 30 days
node
22
Amazon
1
OpenAI (training)
1
Resources
databricks-mcp — pip install databricks-mcp · libregistry