MindfulChat: Mental Health Chatbot

An intelligent, full-stack mental health support chatbot combining BERT-based sentiment analysis (Python) and Google Gemini AI for empathetic, context-aware responses.

---

Features

• BERT-based sentiment analysis for emotional context detection (Python microservice)
• Google Gemini-2.0-flash powered responses (via backend integration)
• Detects multiple emotional states (depression, anxiety, stress, etc.)
• Crisis detection and immediate help resources
• Chat history and sentiment tracking (MongoDB)
• Secure user authentication and JWT-based sessions
• Admin panel for user and chat management
• Modern, responsive web UI

---

🛠 Tech Stack

• Backend: Node.js (Express)
• Frontend: HTML, CSS, JavaScript
• AI/ML:
  • BERT-based sentiment analysis (Python Flask service, HuggingFace Transformers, PyTorch)
  • Google Gemini-2.0-flash (via @google/generative-ai)
• Database: MongoDB (Mongoose ODM)
• Authentication: JWT
• Security: dotenv, rate limiting, CORS, input validation

---

📁 Project Structure

mental_health_chatbot/
├── app.js                  # Main server entry point
├── config/
│   └── db.js               # MongoDB connection
├── controllers/            # Route handlers (chat, user, gemini)
├── middleware/             # Auth/admin JWT middlewares
├── models/                 # Mongoose schemas (User, Chat)
├── public/                 # Frontend (HTML, CSS, JS)
├── routes/                 # Express route definitions
├── sentiment_service/      # Python LSTM sentiment microservice
│   ├── app.py              # Flask API
│   ├── model/              # Model files (tokenizer, vocab, weights)
│   └── requirements.txt    # Python dependencies
├── .env                    # Environment variables (NOT committed)
├── package.json            # Node.js dependencies and scripts
└── README.md

---

⚡️ Quick Start

1. Clone & Install Dependencies

git clone https://github.com/YOUR_USERNAME/mental_health_chatbot.git
cd mental_health_chatbot
npm install

2. Python Sentiment Service Setup

cd sentiment_service
pip install -r requirements.txt
cd ..

3. Environment Variables

Create a .env file in the root directory:

MONGO_URI=your_mongodb_uri
JWT_SECRET=your_jwt_secret
GOOGLE_API_KEY=your_gemini_api_key
PORT=your_port_number
...

4. Run the Application

• Start backend server:

  npm run dev

• Start sentiment service:

  npm run sentiment-service

• Or use the provided PowerShell script:

  npm run start-all

5. Access the App

• Visit [http://localhost:] for the chatbot UI
• Admin panel: [http://localhost:/admin.html]for admin access

---

🔌 API Endpoints (Summary)

Chat

• POST /api/chat/message — Send message to chatbot
• GET /api/chat/history — Get user's chat history

Authentication

• POST /api/auth/register — Register new user
• POST /api/auth/login — Login user

Admin

• GET /api/admin/users — List users
• PATCH /api/admin/user/:id — Edit user
• DELETE /api/admin/user/:id — Delete user
• GET /api/admin/chats — List chats
• ...and more

---

🔒 Security & Privacy

• Sensitive data: All secrets loaded from .env (never committed)
• JWT authentication for all user/admin routes
• Rate limiting and CORS enabled
• Input validation on all endpoints

---

🧠 AI/ML Details

• Sentiment Service: Python Flask app exposes /predict endpoint for BERT-based sentiment analysis using HuggingFace Transformers and PyTorch
• Gemini Integration: Google Gemini API used for generating empathetic, context-aware responses

---

🆘 Crisis Response

• Detects crisis keywords and provides immediate help resources and helplines

---

📄 License

This project is for educational and non-commercial use. Please see LICENSE for details.

---

🙏 Acknowledgments

• Google Gemini for generative AI
• BERT for BERT-based sentiment analysis
• MongoDB for database

---

Note: If you deploy this project, ensure your .env and sensitive files are never pushed to public repositories.

The chatbot includes an automated crisis detection system that:

1. Identifies potential crisis situations
2. Provides immediate helpline information
3. Encourages professional help when needed