Manuka DSIT MCP Demo

A comprehensive demonstration of Model Context Protocol (MCP) server creation and deployment on Databricks, showcasing both Genie Space MCP servers and custom MCP servers for AI-powered data analysis.

Overview

This repository provides end-to-end examples for creating and deploying MCP servers on Databricks, enabling AI agents to access and query organizational data through standardized protocols. The demo includes:

• Genie Space MCP Server: End-to-end pipeline for the creation of a conversational AI interface over structured data (UK Pupil Destinations dataset)
• Custom MCP Server: A purpose-built server wrapping the Department for Transport (DfT) NaPTAN API for transport access point data
• AI Playground Integration: Connect and test MCP servers with Databricks LLMs

This demo is designed for data practitioners, AI engineers, and anyone interested in extending AI capabilities with custom data sources using the Model Context Protocol.

Features

✨ Key Capabilities:

• 🚀 End-to-end MCP server creation examples (Genie Space + Custom)
• 📊 Sample data ingestion and transformation notebooks
• 🏗️ Databricks Apps deployment for custom MCP servers
• 🤖 AI Playground integration for testing and exploration
• 📚 Comprehensive documentation and step-by-step guides

Project Structure

manuka-dsit-mcp-demo/
├── notebooks/                         # Databricks notebooks for data processing
│   ├── 00_set_up.ipynb                # Catalog and schema setup
│   ├── 01_ingest_pupil_dest_data.py   # Data ingestion (Bronze layer)
│   ├── 02a_transform_bronze.ipynb     # CSV to Bronze Delta table
│   ├── 02b_transform_silver.ipynb     # Bronze -> Silver transform
│   └── 03_create_genie_space.py       # Genie Space creation script
│
├── custom_mcps/                       # Custom MCP server implementations
│   └── dft-mcp-server/                # DfT NaPTAN transport data server
│       ├── server/                    # Server source code
│       │   ├── app.py                 # FastAPI + FastMCP setup
│       │   ├── main.py                # Entry point
│       │   ├── tools.py               # MCP tools, resources, prompts
│       │   ├── naptan_api.py          # NaPTAN API client
│       │   └── utils.py               # Utility functions
│       ├── scripts/dev/               # Development scripts
│       ├── tests/                     # Integration tests
│       ├── static/                    # Static web assets
│       ├── pyproject.toml             # Python dependencies
│       ├── app.yaml                   # Databricks Apps config
│       ├── CLAUDE.md                  # AI assistant context
│       └── README.md                  # DfT MCP server documentation
│
├── README.md                          # This file - main documentation
└── .gitignore                         # Git ignore rules

Requirements

Databricks Account

• Databricks workspace (Free Tier or higher)