📘 MindMate — AI Mental Health Chatbot

MindMate is an AI-powered mental health support chatbot that provides empathetic, context-aware responses and emotional support. It detects the user’s emotional state and intensity from text, maintains conversational memory, stores interactions, and visualizes emotional trends through a dashboard.

🌟 Features

✅ Real-time emotion detection (Transformer + Rule-based) ✅ Emotion intensity classification (normal & high) ✅ Session-based conversational memory ✅ Context-aware response generation (avoids repetition) ✅ Crisis detection and safety message delivery ✅ SQLite storage of chat history ✅ Dashboard analytics (emotion counts, trends) ✅ Lightweight deployment on Render ✅ Relaxation tips for emotional coping

🧠 Algorithm Overview

User sends a message

Flask server receives request

Crisis keyword check (safety layer)

Save message to session history (last 5 messages)

Emotion detection (Text → Transformer / Rule)

Emotion intensity classification

Context continuity comparison

Smart response generation

Tip attachment

Save chat & emotion to SQLite

Return response to user

Dashboard updated with new data

This creates a context-aware and emotionally sensitive chatbot that responds meaningfully while tracking emotional trends.

🧩 Tech Stack Component Technology Backend Python, Flask NLP/AI Transformer-based emotion model (optional/prod-light rule) Database SQLite Frontend HTML, CSS, JavaScript Deployment Render Session Flask session for memory 📁 Project Structure MINDMATE/ ├── app.py ├── templates/ │ ├── index.html │ └── dashboard.html ├── static/ ├── requirements.txt ├── mindmate.db ├── emotion_model.py (optional) ├── sentiment_test.py └── emotion_test.py 🚀 Getting Started 1️⃣ Clone the repository git clone https://github.com/Athiraa1804/MINDMATE.git cd MINDMATE 2️⃣ Install dependencies pip install -r requirements.txt 3️⃣ Run the app (local) python app.py

The app will start locally on:

http://127.0.0.1:5000 📊 Dashboard

Navigate to:

http://127.0.0.1:5000/dashboard

Dashboard visualizes: ✔ Total chats ✔ Emotion breakdown ✔ Most frequent emotion

🧠 Emotion Detection

The system supports both modes:

Local fast mode

Rule-based detection

Lightweight

Works on deployment free tiers

Transformer mode (optional)

Install the transformer model in local environment only:

USE_TRANSFORMER=True python app.py

This enables a more accurate emotion classifier.

💾 Database

Chats are stored in:

mindmate.db

Table: chats

Column Type id INTEGER username TEXT message TEXT emotion TEXT timestamp TEXT 📦 Environment Variables

To enable transformer mode in production, configure the following in Render:

USE_TRANSFORMER=True

If not set, the app defaults to lightweight mode.

📍 Crisis Detection

The chatbot detects potentially harmful content like:

suicide, kill myself, die, end my life, can't live

and immediately recommends a helpline.

📌 Deployment (Render)

Push code to GitHub

Connect repo to Render

Set environment variables

Deploy web service

Enjoy online hosted chatbot

Render will automatically build and run your app.

🧩 Future Improvements

✔ Dynamic adaptive learning ✔ Emotion intensity progression ✔ User profile tracking ✔ Advanced analytics graphs ✔ Multi-modal input (voice/image) ✔ Integration with help resources and strategies

📝 License

This project is for academic and educational use.