Metabase MCP Plus

Metabase MCP Plus

An MCP server that enables AI assistants to query databases, execute SQL, and manage Metabase resources like dashboards, cards, and collections through natural language.

Category
访问服务器

README

Metabase MCP Plus

PyPI version Python 3.12+ License: MIT CI

An enhanced Model Context Protocol (MCP) server for Metabase, enabling AI assistants like Claude, Cursor, and other MCP clients to query databases, execute SQL, manage dashboards, cards, and collections — all through natural language.

Quick Start

Using uvx (No Installation Required)

uvx metabase-mcp-plus

Using pip

pip install metabase-mcp-plus
metabase-mcp-plus

From Source

git clone https://github.com/schachan/metabase-mcp.git
cd metabase-mcp
uv sync
uv run python server.py

Configuration

Set environment variables directly or create a .env file:

cp .env.example .env

API Key Authentication (Recommended)

METABASE_URL=https://your-metabase-instance.com
METABASE_API_KEY=your-api-key-here

Email/Password Authentication

METABASE_URL=https://your-metabase-instance.com
METABASE_USER_EMAIL=your-email@example.com
METABASE_PASSWORD=your-password

Optional: Custom Host/Port (SSE/HTTP transports)

HOST=localhost  # Default: 0.0.0.0
PORT=9000       # Default: 8000

Available Tools (22)

Database Operations (6 tools)

Tool Description
list_databases List all configured databases
get_database Get details of a specific database
list_tables Get all tables in a database with metadata
get_table_fields Retrieve field/column information for a table
list_database_schemas List all schemas in a database
get_schema_tables Get tables within a specific schema

Query Operations (1 tool)

Tool Description
execute_query Execute native SQL queries with parameter support

Card / Question Management (6 tools)

Tool Description
list_cards List all saved questions/cards
get_card Get details of a specific card
execute_card Run a saved question and return results
create_card Create a new question with a SQL query
update_card Update an existing card's name, query, or collection
archive_card Archive a card

Dashboard Management (5 tools)

Tool Description
list_dashboards List all dashboards
get_dashboard Get full dashboard details including cards
create_dashboard Create a new dashboard
update_dashboard Update a dashboard's name, description, or cards
delete_dashboard Delete a dashboard

Collection Management (3 tools)

Tool Description
list_collections Browse all collections
get_collection_items List items within a collection
create_collection Create a new collection

Search (1 tool)

Tool Description
search Search across cards, dashboards, collections, and tables

Transport Methods

metabase-mcp-plus                # STDIO (default, for IDE integration)
metabase-mcp-plus --sse          # Server-Sent Events
metabase-mcp-plus --http         # Streamable HTTP

Or from source:

uv run python server.py          # STDIO
uv run python server.py --sse    # SSE
uv run python server.py --http   # HTTP

IDE Integration

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
    "mcpServers": {
        "metabase": {
            "command": "uvx",
            "args": ["metabase-mcp-plus"],
            "env": {
                "METABASE_URL": "https://your-metabase-instance.com",
                "METABASE_API_KEY": "your-api-key-here"
            }
        }
    }
}

Cursor

Add to your MCP settings (.cursor/mcp.json):

{
    "mcpServers": {
        "metabase": {
            "command": "uvx",
            "args": ["metabase-mcp-plus"],
            "env": {
                "METABASE_URL": "https://your-metabase-instance.com",
                "METABASE_API_KEY": "your-api-key-here"
            }
        }
    }
}

Project Structure

metabase-mcp/
├── server.py              # Entry point — CLI arg parsing and transport selection
├── app.py                 # FastMCP app setup, env loading, client instantiation
├── client.py              # MetabaseClient with auth, retry, and error handling
├── tools/
│   ├── __init__.py        # Registers all tool modules
│   ├── databases.py       # Database, table, schema, and field tools
│   ├── queries.py         # SQL query execution
│   ├── cards.py           # Card CRUD and execution
│   ├── dashboards.py      # Dashboard CRUD
│   ├── collections.py     # Collection browsing and creation
│   └── search.py          # Cross-entity search
├── tests/                 # Comprehensive test suite (95% coverage)
├── .github/workflows/
│   ├── ci.yml             # Lint + test on push/PR
│   └── publish.yml        # Release → PyPI + attach assets
├── RELEASING.md           # How to cut a release (trusted publisher setup)
└── pyproject.toml

Development

# Install with dev dependencies
uv sync --group dev

# Run linting and formatting
uv run ruff check .
uv run ruff format .

# Run tests with coverage
uv run pytest tests/ --cov=. --cov=tools --cov-report=term-missing -v

Architecture Highlights

  • Modular tools — each domain (databases, cards, dashboards, etc.) is a separate module
  • Automatic auth retry — expired session tokens are refreshed transparently on 401
  • Custom exceptionsMetabaseAPIError and MetabaseAuthError for typed error handling
  • Error handler decorator@tool_error_handler eliminates boilerplate try/except across all 22 tools
  • Middleware stack — built-in error handling and logging middleware via FastMCP

Releases & PyPI

Publishing to PyPI is automated when you publish a GitHub Release (workflow .github/workflows/publish.yml): lint → test → build → attach wheels to the release → upload to PyPI (trusted publisher / OIDC).

See RELEASING.md for one-time PyPI + GitHub environment setup and the exact tag/version rules (v1.2.3 must match version = "1.2.3" in pyproject.toml).

License

MIT License — see LICENSE for details.

Resources

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选