Metabase MCP Server

Metabase MCP Server is an MCP server that connects Large Language Models to the Metabase business intelligence platform via its official API. Data analysts, engineers, and product teams use this server to interact directly with internal reporting systems, dashboards, and connected databases from their MCP-compatible AI clients. The integration enables assistants to inspect organizational data models by listing available databases, collections, and existing Metabase cards. Users can ask their AI client to execute queries associated with existing cards or run arbitrary SQL and database queries directly against the configured Metabase instance. In addition to reading and exploring business data, the server allows clients to create new cards and add bookmarks, facilitating conversational report creation and insight capture. Running locally using Python 3.12 and the uv package manager, the server bridges the gap between conversational AI workflows and production business intelligence repositories without requiring users to switch between browser tabs or manually export data sheets.

Category: CRM, ERP & E-commerce

Tags: business-intelligence, dashboards, data-visualization, metabase, SQL

Visit Metabase MCP Server

How to install and configure Metabase MCP Server

1. Clone the repository and navigate to the folder: bash git clone https://github.com/cuenca-mx/metabase-mcp-server.git cd metabase-mcp-server 2. Ensure Python 3.12+ and uv are installed. Run make install for local setup. 3. Add the server to your MCP client configuration (such as Claude Desktop or Cursor): json { "mcpServers": { "metabase": { "command": "uv", "args": [ "--directory", "/ABSOLUTE/PATH/TO/PARENT/FOLDER/metabase_mcp_server", "run", "app.py" ], "env": { "METABASE_URL": "https://metabase.domain.com/", "METABASE_API_KEY": "your-secret-api-key" } } } } 4. Update the path to match your local installation and set your valid METABASE_URL and METABASE_API_KEY.

What you can do with Metabase MCP Server

• Querying existing Metabase cards directly through an LLM to retrieve fresh KPI metrics and analytical tables for reporting. - Executing arbitrary ad hoc SQL queries against connected enterprise databases to answer specific operational questions without opening the Metabase interface. - Browsing organization collections and database schemas to help engineers and analysts locate specific data assets and card definitions. - Creating new Metabase visualization cards and bookmarks directly from conversational AI prompts to speed up dashboard construction workflows.

Key facts

• Open Source
• https://github.com/cuenca-mx/metabase-mcp-server
• CRM, ERP & E-commerce, Data & Analytics
• business-intelligence, dashboards, data-visualization, metabase, SQL

Part of MCP Servers

What can Metabase MCP Server do?

Metabase MCP Server allows AI clients to interact with Metabase deployments. It exposes tools to list available databases, list collections, inspect Metabase cards, execute queries stored on cards, run arbitrary database queries, generate new cards, and create bookmarks.

How do I install Metabase MCP Server?

You install the server by cloning the GitHub repository and using the uv package runner on Python 3.12 or higher. Add the server entry to your client configuration file with the command uv and arguments pointing to app.py, alongside your Metabase URL and API key.

Which MCP clients work with Metabase MCP Server?

Any client supporting the Model Context Protocol can connect to Metabase MCP Server, including Claude Desktop, Cursor, and debugging utilities like the MCP Inspector using standard stdio communication.

Is Metabase MCP Server open source?

Yes, Metabase MCP Server is an open-source project hosted on GitHub by cuenca-mx. Anyone can inspect the code, contribute improvements, run unit tests with pytest, and deploy it locally.