MCP.so
ログイン

mcp-server-unitycatalog: An Unity Catalog MCP server

@ognis1205

リポジトリ

mcp-server-unitycatalog: An Unity Catalog MCP server について

Unity Catalog AI Model Context Protocol Server

#mcp#llm#unitycatalog#modelcontextprotocol

基本情報

カテゴリ

その他

ライセンス

MIT

ランタイム

python

トランスポート

stdio

公開者

ognis1205

設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "mcp-server-unitycatalog": {
      "command": "docker",
      "args": [
        "build",
        "-t",
        "mcp/unitycatalog",
        "."
      ]
    }
  }
}

ツール

7

A list of functions retrieved from Unity Catalog.

`name` (string): The name of the function (not fully-qualified).

A function details retrieved from Unity Catalog.

`name` (string): The name of the function (not fully-qualified).

A function details created within Unity Catalog.

`name` (string): The name of the function (not fully-qualified).

None.

概要

What is mcp-server-unitycatalog?

A Model Context Protocol (MCP) server that exposes Unity Catalog Functions as MCP tools. It allows AI assistants to list, retrieve, create, and delete functions registered in a Unity Catalog server. This server is intended for developers using Unity Catalog who want to manage functions through natural language or AI interfaces.

How to use mcp-server-unitycatalog?

Run directly with uvx (no installation) or via Docker. Configure the server using CLI arguments or environment variables: the Unity Catalog server URL (--uc_server), catalog (--uc_catalog), schema (--uc_schema) are required; an access token (--uc_token), verbosity level, and log directory are optional. Add the server to Claude Desktop or VSCode Cline by including a JSON entry in their MCP settings files.

Key features of mcp-server-unitycatalog

  • Provides Unity Catalog Functions as MCP tools.
  • uc_list_functions: lists functions within a catalog/schema.
  • uc_get_function: retrieves details of a specific function.
  • uc_create_function: creates a new Python function (experimental).
  • uc_delete_function: deletes a function from Unity Catalog.
  • Runs via uv or Docker with no permanent install needed.

Use cases of mcp-server-unitycatalog

  • AI assistants querying available Unity Catalog functions.
  • Retrieving function signatures and metadata interactively.
  • Creating new Python functions in Unity Catalog through chat.
  • Deleting obsolete functions with natural language commands.
  • Integrating Unity Catalog function management into Claude Desktop or VSCode.

FAQ from mcp-server-unitycatalog

What is the difference between mcp-server-unitycatalog and Unity Catalog's native API?

This server acts as an MCP bridge, making Unity Catalog Functions available directly to AI models via standard MCP tools, rather than requiring direct REST API calls.

Are there any runtime dependencies?

The server requires Python and uv (or Docker) to run. No additional dependencies are installed permanently when using uvx.

Where does function data live?

All function data resides in the remote Unity Catalog server. The MCP server is a client that queries and modifies that data.

Is the uc_create_function tool stable?

No, it is marked experimental and may change in future versions.

What transport and authentication are used?

The server communicates over stdio (MCP transport). Authentication is via an optional access token passed as a CLI argument or environment variable.

コメント

「その他」の他のコンテンツ

Sequential Thinking

modelcontextprotocol

Model Context Protocol Servers

88.2K

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

Website

FunnyWolf

Adversary simulation and Red teaming platform with AI

5.1K

Mobile Mcp

mobile-next

Model Context Protocol Server for Mobile Automation and Scraping (iOS, Android, Emulators, Simulators and Real Devices)

5.4K

ACI: Open-Source Infra to Power Unified MCP Servers

aipotheosis-labs

ACI.dev is the open source tool-calling platform that hooks up 600+ tools into any agentic IDE or custom AI agent through direct function calling or a unified MCP server. The birthplace of VibeOps.

4.8K