AI Customer Support Chatbot Platform

A modular, production-ready AI-powered customer support chatbot platform that handles multi-channel communication via WhatsApp, web chat, and email.

🧠 Core Features

AI Chatbot with GPT-based Intelligence

• OpenAI GPT-4o for text responses
• Whisper for audio transcriptions (voice messages)
• Optional DALL·E or CLIP for image generation/understanding

Multi-Channel Support

• WhatsApp integration using WhatsApp Cloud API / Twilio
• Web chat widget embeddable on websites (React component)
• Email-based chat processing

Contextual Memory

• Per-user, multi-session context across platforms using Redis
• Long-term memory storage with ChromaDB for embeddings and search

Multilingual Support

• Detect and respond in user's language using GPT-4o
• Optional translation layer for fallback

Live-Agent Handover

• Seamless switch from AI to human agent
• Agent notification system with chat history
• Real-time or asynchronous takeover

Admin Dashboard

• Built with Next.js + TailwindCSS
• Role-based access for chat management and system configuration

🏗️ Project Structure

ai-customer-support-platform/
├── services/
│   ├── chatbot-core/           # Core chatbot service
│   ├── memory-engine/          # Memory and context management
│   ├── channel-integrations/   # WhatsApp, Web, Email integrations
│   ├── live-agent-engine/      # Agent handover system
│   └── admin-api/              # API for admin dashboard
├── frontend/
│   ├── admin-dashboard/        # Next.js admin dashboard
│   └── chat-widget/            # Embeddable chat widget
├── infrastructure/
│   ├── docker/                 # Docker configurations
│   └── kubernetes/             # Kubernetes deployment files
└── docs/                       # Documentation

📦 Getting Started

Prerequisites

• Node.js 18+
• Python 3.9+
• Docker and Docker Compose
• OpenAI API key

Installation

1. Clone the repository
2. Set up environment variables
3. Run docker-compose up to start all services

🚀 Development

See individual service READMEs for specific development instructions.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🔌 Client Libraries

Memory Engine Client Libraries

The Memory Engine service provides client libraries in multiple languages for seamless integration with other components:

JavaScript/TypeScript Client

• Full-featured client for Node.js and browser applications
• Strongly-typed TypeScript interface with complete type definitions
• Supports all Memory Engine endpoints with proper error handling
• Includes timeout handling and comprehensive documentation

Python Client

• Asynchronous and synchronous client implementations
• Complete support for all Memory Engine features
• Proper error handling and logging

Features Supported by All Clients

• Memory management (short-term and long-term)
• Knowledge base semantic search
• Vector embedding creation
• Session context management
• Context optimization for token efficiency
• Health checks

Usage Examples

Each client library includes comprehensive examples demonstrating all features:

• Node.js example
• TypeScript example
• Python example (async and sync)

See the memory-engine/client/README.md for detailed documentation, installation instructions, and API reference.