MindDecode - HackThrone project

Bridging Academic Intelligence and Emotional Resilience through AI

Live Demo:
🎯 https://prep-nova-minddecode.replit.app/

---

🧩 Project Overview

PrepNOVA – MindDecode is an AI-powered academic wellness and mock test platform that empowers students to achieve balance between academic performance and mental well-being.

Developed for a hackathon, it fuses AI, Machine Learning (ML), and EdTech innovation to deliver:

• Personalized study recommendations
• Cognitive wellness insights
• Adaptive mock testing
• Emotion-aware learning analytics

---

🌟 Key Features

🤖 AI Chat Guidance

GPT-integrated chatbot for academic and emotional support, providing tailored study help and motivational nudges.

---

📈 ML-Based Performance Predictor

A custom ML model built with scikit-learn to analyze:

• Past test scores
• Difficulty levels of subjects
• Cognitive load and stress trends

It then predicts focus areas for improvement.

---

🧠 Dynamic Mock Test Platform

• AI-curated mock tests
• Instant feedback
• Adaptive question difficulty
• Progress tracking

---

📊 Performance Dashboard

Interactive visual dashboard that displays:

• Topic-wise accuracy and performance
• Progress over time
• Personalized improvement suggestions

---

🔐 Secure Authentication

Powered by Clerk, ensuring:

• Encrypted sign-in / sign-up
• JWT-based user sessions
• Data privacy and protection

---

☁️ Cloud & Replit Deployment

Deployed entirely on Replit, ensuring seamless accessibility and zero-setup execution.

---

🧠 Tech Stack

Layer	Technology
Frontend	React + TypeScript (Vite)
Backend	Node.js + Express + TypeScript
AI/ML	OpenAI GPT API + Custom ML Predictor (Python)
Database	PostgreSQL / SQLite
Authentication	Clerk
Deployment	Replit Cloud

---

🏗️ Project Structure

PrepNOVA/
├── server/
│   ├── index.ts
│   ├── routes/
│   ├── ml/
│   └── middleware/
├── frontend/
│   ├── src/
│   │   ├── pages/
│   │   ├── components/
│   │   └── App.tsx
├── data/
│   └── db.sqlite
├── .env
└── README.md

---

🔒 Security Implementation

• 🔑 Clerk Authentication for secure user access
• 🌐 HTTPS & Secure CORS policies
• 🧩 Encrypted environment variables for API keys
• 🧼 Database sanitization to prevent injection attacks

---

⚙️ Workflow

1. 🧍 User logs in securely via Replit Authentication
2. 🧭 Chooses between Mock Test or MindDecode (Predictor) mode
3. 🧮 Backend processes test or prediction requests
4. 🧠 ML module generates insights and improvement focus areas
5. 📊 Dashboard visualizes performance metrics and recommendations

---

🧪 Machine Learning Model

The MindDecode Predictor model evaluates:

• 🧾 Past academic scores and metrics
• 🧠 Difficulty correlation between topics
• 💭 Cognitive load and stress patterns

Built With:

• Python
• scikit-learn
• pandas / numpy
• Flask (REST API integration)

---

---

🔒 Security Implementation

• 🔑 Clerk Authentication for secure user access
• 🌐 HTTPS & Secure CORS policies
• 🧩 Encrypted environment variables for API keys
• 🧼 Database sanitization to prevent injection attacks

---

⚙️ Workflow

1. 🧍 User logs in securely via Replit Authentication
2. 🧭 Chooses between Mock Test or MindDecode (Predictor) mode
3. 🧮 Backend processes test or prediction requests
4. 🧠 ML module generates insights and improvement focus areas
5. 📊 Dashboard visualizes performance metrics and recommendations

---

🧪 Machine Learning Model

The MindDecode Predictor model evaluates:

• 🧾 Past academic scores and metrics
• 🧠 Difficulty correlation between topics
• 💭 Cognitive load and stress patterns

Built With:

• Python
• scikit-learn
• pandas / numpy
• Flask (REST API integration)

---

⚡ Installation & Setup

1️⃣ Clone the Repository

git clone https://github.com/aishwarya-mol-2046/mind_decode.git
cd mind_decode

📦 Installation & Setup

2️⃣ Install Dependencies

npm install

3️⃣ Configure Environment Variables

Create a .env file and add:

MODEL=auto
OPENAI_API_KEY=your_openai_key_here
GEMINI_API_KEY=optional_google_api_key
DATABASE_URL=postgresql://postgres:password@localhost:5432/prepnova

4️⃣ Run the App

npm run dev

---

Screenshots of the output

Home page

dashboard page

AI questions predictor page

1. Upload the files for the needed subject

2. Predict the question paper for the given subject

Promodoro Timer

1. Start the timer

2. Stop when you feel out of focus and give your feedback 3. Click submit and continue to get your mood analysis

Mock up test

1. Attend the questions

2. View result analysis, also have an option for retest

---

📜 Author

Niveditha Venkatesh

---

📜 License

MIT License Copyright (c) 2025 Niveditha Venkatesh

---