WhatsApp LLM Chatbot: Your Privacy-First AI Assistant

Transform your WhatsApp experience with a locally-deployed intelligent AI powered by DeepSeek-R1, ensuring 100% data privacy while delivering state-of-the-art conversational intelligence.

WhatsApp Chat Interface

🔒 Privacy-First Architecture

• Local LLM Deployment: All processing happens on your hardware - no data leaves your system
• Zero Cloud Dependencies: No connections to DeepSeek servers or third-party services
• Complete Data Control: You own and control all conversation data
• Open Source Transparency: Fully auditable codebase

🧠 Powered by DeepSeek-R1

Our solution leverages the latest DeepSeek-R1 model, offering:

• Advanced natural language understanding
• Context-aware responses
• Multi-turn conversation capabilities
• Customizable knowledge domains based on your data input

🌟 Vision & Use Cases

Transform your business communications with privacy-preserving AI technology. This versatile solution enables:

Customer Service

• 24/7 instant customer support
• Intelligent query handling
• Automated ticket creation and tracking
• Multi-language support

Business Operations

• Meeting scheduling and calendar management
• Order tracking and updates
• Product recommendations
• Invoice generation and processing

Educational Support

• Interactive tutoring
• Course information delivery
• Assignment assistance
• Study planning

Healthcare Assistance

• Appointment scheduling
• Medication reminders
• Basic health inquiries
• Wellness tips and tracking

Personal Assistant

• Task management
• Event planning
• Travel booking assistance
• Restaurant recommendations

More Use Cases

🚀 Features

• Real-time conversation with local LLM models
• Webhook integration for instant message processing
• Media message support (images, documents, audio)
• Encrypted media handling
• Message status tracking

🛠️ Technical Stack

• Node.js
• Express.js
• Ollama (Local LLM)
• WhatsApp Business API
• Webhook Integration

🔧 Installation & Setup

Prerequisites

• Node.js installed
• Ngrok account
• Meta Developer account
• Ollama installed locally

Clone the Repository

git clone https://github.com/yourusername/whatsapp_ollama_bot.git
cd whatsapp_ollama_bot

Install Dependencies

npm install express axios ollama-node dotenv

Configure Environment Variables

Create a .env file in the root directory with the following content:

WHATSAPP_TOKEN=your_whatsapp_token
VERIFY_TOKEN=your_verify_token

Set Up Ngrok

1. Download Ngrok from the official website.
2. Extract and run:

   ./ngrok http 5001

3. Copy the generated HTTPS URL.

Configure Meta Webhook

1. Go to the Meta Developers Portal.
2. Navigate to WhatsApp > Configuration.
3. Set Callback URL to: your_ngrok_url/webhook
4. Enter your Verify Token.

Set Up Ollama

1. Start the Ollama service:

   ollama serve

2. Pull and run the DeepSeek-R1 model:

   ollama run deepseek-r1

3. Verify the model is running:

   ollama ps

Start the Application

node index.js

---

📋 Development Roadmap

Current Features

• Real-time WhatsApp message processing
• Local LLM integration
• Basic conversation handling
• Webhook implementation

Upcoming Features

• Web search integration
• Vector-DB/MongoDB for RAG search
• Multi-modal processing (image/video/voice)
• Enhanced RAG capabilities
• Performance optimization
• Admin dashboard

---

📊 Performance Optimization

Hardware Requirements

• GPU: 12GB+ VRAM recommended
• Optimal performance with RTX 4070/4080/4090
• M1/M2 Macs with 16GB+ RAM supported

Model Selection

• Currently using deepseek-r1 7B model
• Optimized for performance/quality balance

Performance Metrics

---

📚 Documentation

• WhatsApp Cloud API Documentation
• Ollama Documentation
• DeepSeek-R1 Model Details

---

📄 License

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