Manage Microsoft Fabric with PySpark & LLMs
MCP server exposing Microsoft Fabric API tools for managing workspaces, lakehouses, warehouses, and PySpark notebooks with intelligent code generation and
Maintainer of this project? Claim this page to edit the listing.
1.0.0Add to Favorites
Why it matters
Streamline your Microsoft Fabric development and MLOps with this comprehensive Python MCP server. It integrates LLMs for intelligent notebook creation, code generation, and performance optimization.
Outcomes
What it gets done
Develop and optimize PySpark notebooks with LLM assistance.
Manage Fabric resources like workspaces, lakehouses, and warehouses.
Automate data loading, querying, and schema management.
Analyze and improve notebook performance with detailed recommendations.
Install
Add it to your toolbox
Run in your project directory:
curl -fsSL https://spark.entire.vc/get/vb-fabric-mcp | bash Capabilities
Tools your agent gets
Show all available Fabric workspaces
Set the current workspace context for the session
Show all lakehouses in the workspace
Create a new lakehouse
Set the current lakehouse context
Show all warehouses in the workspace
Create a new warehouse
Set the current warehouse context
Overview
Fabric MCP Server
A Python-based MCP server that connects AI assistants to Microsoft Fabric APIs for workspace, lakehouse, warehouse, and table management. Provides tools for executing SQL queries, loading data, and creating PySpark notebooks with specialized templates and performance optimization features. Use when you need to automate Microsoft Fabric operations through an AI assistant, develop and optimize PySpark notebooks, or manage Fabric resources programmatically. Ideal for data engineers working with Fabric workspaces who want conversational access to API operations.
What it does
Big Job: Accelerate Microsoft Fabric data engineering workflows by giving AI assistants direct access to workspace management, SQL execution, and intelligent PySpark notebook development.
Small Job: Connect your AI client to Fabric APIs so you can ask "List all my Fabric workspaces" or "Create a PySpark notebook that reads sales data, cleans it, and optimizes performance" and get immediate results. Includes 6 specialized notebook templates (basic, ETL, analytics, ML, fabric_integration, streaming), performance scoring (0-100), and optimization recommendations.
Setup (STDIO):
{
"mcp": {
"servers": {
"ms-fabric-mcp": {
"type": "stdio",
"command": "<FullPathToProjectFolder>\\.venv\\Scripts\\python.exe",
"args": ["<FullPathToProjectFolder>\\fabric_mcp.py"]
}
}
}
}
Supports both STDIO and HTTP modes. Requires Azure authentication (az login --scope https://api.fabric.microsoft.com/.default).
Source README
A comprehensive Python-based MCP server for working with Microsoft Fabric API, featuring advanced capabilities for developing, testing, and optimizing PySpark notebooks with LLM integration.
Installation
From Source
git clone https://github.com/your-repo/fabric-mcp.git
cd fabric-mcp
uv sync
pip install -r requirements.txt
MCP Inspector
uv run --with mcp mcp dev fabric_mcp.py
HTTP Server
uv run python .\fabric_mcp.py --port 8081
Configuration
VSCode STDIO Integration
{
"mcp": {
"servers": {
"ms-fabric-mcp": {
"type": "stdio",
"command": "<FullPathToProjectFolder>\\.venv\\Scripts\\python.exe",
"args": ["<FullPathToProjectFolder>\\fabric_mcp.py"]
}
}
}
}
VSCode HTTP Integration
{
"mcp": {
"servers": {
"ms-fabric-mcp": {
"type": "http",
"url": "http://<localhost or remote IP>:8081/mcp/",
"headers": {
"Accept": "application/json,text/event-stream"
}
}
}
}
}
Available Tools
| Tool | Description |
|---|---|
list_workspaces |
Show all available Fabric workspaces |
set_workspace |
Set the current workspace context for the session |
list_lakehouses |
Show all lakehouses in the workspace |
create_lakehouse |
Create a new lakehouse |
set_lakehouse |
Set the current lakehouse context |
list_warehouses |
Show all warehouses in the workspace |
create_warehouse |
Create a new warehouse |
set_warehouse |
Set the current warehouse context |
list_tables |
Show all tables in the lakehouse |
get_lakehouse_table_schema |
Get the schema for a specific table |
get_all_lakehouse_schemas |
Get schemas for all tables in the lakehouse |
set_table |
Set the current table context |
get_sql_endpoint |
Get the SQL endpoint for a lakehouse or warehouse |
run_query |
Execute SQL queries |
load_data_from_url |
Load data from a URL into tables |
Capabilities
- Manage workspaces, lakehouses, warehouses, and tables
- Retrieve Delta table schemas and metadata
- Execute SQL queries and load data
- Perform operations on reports and semantic models
- Intelligent notebook creation with 6 specialized templates
- Smart code generation for common PySpark operations
- Comprehensive validation with syntax checking and best practices
- Fabric-specific optimizations and compatibility checks
- Performance analysis with scoring and optimization recommendations
- Real-time monitoring and execution analytics
Usage Examples
List all my Fabric workspaces
Create a PySpark notebook that reads sales data, cleans it, and optimizes performance
My PySpark notebook is slow. Help me optimize it.
Notes
Requires Azure authentication (az login --scope https://api.fabric.microsoft.com/.default). Includes 6 specialized PySpark templates: basic, etl, analytics, ml, fabric_integration, and streaming. Provides performance scoring (0-100) and detailed optimization recommendations. Supports both STDIO and HTTP communication modes.
FAQ
Common questions
Discussion
Questions & comments · 0
Sign In Sign in to leave a comment.