ms-fabric-mcp

# MS Fabric MCP Server This project provides a Model Context Protocol (MCP) server that enables clients to query and explore schemas in Microosft Fabric items. The goal is to enable more robust automation of data engineering flows by enabling the agent/llm to query and verify Enable LLMs to query lakehouses, warehouses and SQL databases to make automation of building data pipelines mor robust. It uses Azure Active Directory (AAD) token authentication. ## Prerequisites * **Python 3.10+** installed. * **Microsoft ODBC Driver for SQL Server** installed. * https://learn.microsoft.com/en-us/sql/connect/odbc/download-odbc-driver-for-sql-server?view=sql-server-ver16 * Commonly "ODBC Driver 17 for SQL Server" or "ODBC Driver 18 for SQL Server". Driver 18 is recommended for better AAD support. * **Azure CLI** installed and authenticated: * Run `az login` and complete the authentication flow, `az login --allow-no-subscriptions` might be necessary. * The authenticated user/principal must have appropriate permissions (e.g., `contributor` role in the workspace or owner of the Fabric item). ## Setup Instructions pip installation ```bash pip install ms-fabric-mcp ``` uv installation ```bash uv add ms-fabric-mcp # or uv pip install ms-fabric-mcp ``` Example `mcp.json` configuration for use with Cursor, Claude Desktop, etc. ```json "ms-fabric-mcp": { "command": "uv", "args": [ "--directory", "path/to/server", "run", "mcp" ], "env": { "SQL_SERVER_NAME": "xyz-xyz.datawarehouse.fabric.microsoft.com", "SQL_DATABASE_NAME": "dev", "ODBC_DRIVER": "{ODBC Driver 18 for SQL Server}" } } ``` ### Alternatives 1. **Clone the Repository (Optional):** ```bash git clone <repository_url> cd ms_fabric_mcp # Or your project directory name ``` 2. **Install Dependencies:** ```bash uv sync ``` provide required env varia…