customer-copilot-app

tylerwatson-db/customer-copilot-app

3.1

If you are the rightful owner of customer-copilot-app and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

The Databricks MCP Server Template enables hosting Model Context Protocol (MCP) prompts and tools on Databricks Apps, allowing AI assistants like Claude to interact with your Databricks workspace securely.

Tools
2
Resources
0
Prompts
0

🤖 Customer Copilot App

A clean, modern web application for querying a customer copilot AI agent deployed on Databricks model serving.

What is this?

This is a streamlined customer copilot application that provides:

  • 📝 Clean web interface for querying the customer copilot model
  • 🔐 Secure OAuth authentication through Databricks Apps
  • 🚀 Instant deployment to Databricks Apps
  • ⚡ Real-time model queries to the customer copilot agent

Features

Customer Copilot Query Interface

  • Clean, minimalist web UI
  • Real-time query processing
  • Response time tracking
  • Copy-to-clipboard functionality
  • Error handling and display

Model Serving Integration

  • Connects to Databricks model serving endpoint
  • OAuth authentication via Databricks SDK
  • Automatic token refresh
  • Timeout handling and retry logic

Quick Start

Deploy to Databricks

# Clone the repository
git clone https://github.com/tylerwatson-db/customer-copilot-app.git
cd customer-copilot-app

# Run the interactive setup
./setup.sh

# Deploy to Databricks Apps
./deploy.sh

Local Development

# Install dependencies
uv sync

# Start the development server
./run_app_local.sh

Architecture

┌─────────────┐    HTTP/JSON    ┌──────────────────┐    OAuth    ┌─────────────────┐
│   Web UI    │ ◄─────────────► │  FastAPI Server  │ ◄─────────► │ Databricks App  │
│ (React/TS)  │                 │   (Python)       │             │ (Model Serving) │
└─────────────┘                 └──────────────────┘             └─────────────────┘

Components

  1. Frontend (client/): React TypeScript application with:

    • QueryInterface component for user interaction
    • Modern UI with Tailwind CSS
    • Real-time query processing

  2. Backend (server/): FastAPI application with:

    • /api/query endpoint for model queries
    • OAuth authentication via Databricks SDK
    • Static file serving for the React app

  3. Model Integration: Connects to customer copilot model serving endpoint

Configuration

The app is configured via:

  • .env.local: Databricks authentication and app settings
  • config.yaml: App configuration
  • app.yaml: Databricks App deployment configuration

Deployment

The app is deployed to Databricks Apps and accessible at:

  • App URL: https://customer-copilot-app-475848639457152.aws.databricksapps.com
  • Workspace URL: https://fe-vm-team-nasty-hackathon-ws.cloud.databricks.com/apps/customer-copilot-app

API Usage

Query the Customer Copilot

POST /api/query
Content-Type: application/json

{
  "query": "Your customer question here"
}

Response Format

{
  "response": "AI response from the customer copilot",
  "metadata": {
    "responseTime": 1.23,
    "timestamp": "2025-08-29T13:18:21Z",
    "endpoint": "https://..."
  },
  "toolsUsed": ["sql", "search", "databricks"],
  "error": null
}

Development

Project Structure

├── client/                 # React TypeScript frontend
│   ├── src/
│   │   ├── components/     # UI components
│   │   │   └── QueryInterface.tsx  # Main query interface
│   │   └── App.tsx         # Main app component
│   └── build/              # Production build
├── server/                 # FastAPI backend
│   ├── routers/            # API routes
│   │   ├── query.py        # Query endpoint
│   │   └── user.py         # User info endpoint
│   └── app.py              # Main FastAPI app
├── deploy.sh               # Deployment script
├── setup.sh                # Setup script
└── app.yaml                # Databricks App config

Available Scripts

  • ./setup.sh - Interactive setup and configuration
  • ./deploy.sh - Deploy to Databricks Apps
  • ./run_app_local.sh - Run locally for development
  • ./app_status.sh - Check app status and get URLs

License

See LICENSE.md