Premium Dental Practice Chatbot

A sophisticated, scalable conversational AI chatbot for dental practices built with LangGraph, LangChain, and modern web technologies. This solution provides natural, human-like interactions for patient management, appointment scheduling, and practice inquiries.

🏗️ Architecture

This project follows Clean Architecture and SOLID principles for maximum scalability and maintainability:

├── domain/           # Core business logic and entities
├── application/      # Use cases and orchestration
├── infrastructure/   # External integrations (LLM, database)
└── presentation/     # API and frontend

Key Technologies

• LangGraph: State machine for conversation workflows
• LangChain: LLM integration and tool calling
• FastAPI: High-performance async backend
• React/Next.js: Premium frontend UI
• Pydantic: Type-safe data validation

🚀 Quick Start

Prerequisites

• Python 3.10+
• Node.js 18+ (for frontend)
• Free-tier API key from one of:
  • DeepSeek
  • Google Gemini
  • OpenAI

Setup

1. Clone and navigate to the project:

cd "Dentist Coversational AI"

2. Create virtual environment:

python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Configure environment:

cp .env.template .env
# Edit .env and add your API key

5. Initialize database:

mkdir -p data
python scripts/init_database.py

6. Start backend:

uvicorn main:app --reload --host 0.0.0.0 --port 8000

7. Start frontend (in another terminal):

cd frontend
npm install
npm run dev

8. Access the application:

• Frontend: http://localhost:3000
• API Docs: http://localhost:8000/docs

📋 Features

Core Capabilities

1. New Patient Registration

   • Collects: Full name, phone, date of birth, insurance
   • Guides through first appointment booking

2. Existing Patient Management

   • Identity verification
   • Appointment rescheduling
   • Appointment cancellation

3. Intelligent Appointment Scheduling

   • Handles subjective dates ("later next week", "early next month")
   • Suggests alternatives when times don't work
   • Supports multiple appointment types:
     • Cleaning
     • General checkup
     • Emergency (with staff notification)

4. Family Scheduling

   • Book multiple appointments for family members
   • Coordinate back-to-back appointments
   • Manage family member relationships

5. General Inquiries

   • Insurance and payment options
   • Self-pay and membership plans
   • Location and hours information

6. Emergency Handling

   • Captures emergency details
   • Notifies staff immediately
   • Provides appropriate guidance

🎨 Design Philosophy

• Premium UI: Clean, Apple-inspired design with minimal clutter
• Natural Conversations: Human-like, context-aware responses
• Scalable Architecture: SOLID principles, dependency injection, clean separation
• Error Resilience: Comprehensive fallback mechanisms

📁 Project Structure

.
├── domain/              # Domain models and business logic
│   ├── entities/        # Core entities (Patient, Appointment, etc.)
│   └── value_objects/   # Value objects
├── application/         # Application layer
│   ├── use_cases/       # Business use cases
│   └── services/       # Application services
├── infrastructure/      # Infrastructure layer
│   ├── llm/            # LLM integrations
│   ├── database/       # Database implementations
│   └── tools/          # LangChain tools
├── presentation/        # Presentation layer
│   ├── api/            # FastAPI routes
│   └── frontend/       # React frontend
├── data/               # Database files
├── scripts/            # Utility scripts
└── main.py            # Application entry point

🔧 Configuration

All configuration is managed through environment variables (see .env.template). Key settings:

• LLM Provider: Choose DeepSeek, Gemini, or OpenAI
• Database: Local JSON (default) or Supabase
• Practice Info: Customize practice name, hours, contact

🧪 Testing Scenarios

The chatbot handles these key scenarios:

1. New patient booking a cleaning appointment
2. Existing patient rescheduling due to conflict
3. Family booking (parent + 2 kids, back-to-back)
4. Emergency appointment with details
5. Insurance inquiry for non-insured patient
6. Subjective date parsing ("next Tuesday afternoon")

📝 API Documentation

Interactive API documentation available at /docs when the server is running.

🎥 Demo Video

[Link to demo video will be added]

🤝 Contributing

This is an assessment project demonstrating:

• Rapid development capabilities
• Clean architecture principles
• Production-ready code quality
• Comprehensive error handling
• User experience excellence

📄 License

Assessment Project - Internal Use