🎮 GameGuard AI

Detecting unfair gameplay using intelligent telemetry

GameGuard AI is a full-stack machine learning application built to detect cheating behavior in online multiplayer games. It analyzes gameplay telemetry, system info, and behavioral patterns to flag potential cheaters using an intelligent model.

---

🚀 Features

• 📄 Upload game telemetry data as CSV
• 🧠 Detect cheaters based on system + gameplay signals
• 📊 View predictions with confidence levels
• 🔐 Flags like "VM Detected" and "Secure Boot Off"
• 🌙 Light/Dark mode toggle
• 🔄 Real-time feedback with loading indicators
• 📜 Scrollable prediction result table

---

🧱 Tech Stack

Frontend	Backend	ML/Infra
React + Vite + TailwindCSS	Flask + Python	Scikit-learn, Pandas, Joblib

---

📂 Folder Structure

GameGuardAI/
├── backend/
│   ├── app.py                  # Flask server
│   └── models/                 # ML model + utilities
│       ├── model.py
│       ├── *.pkl               # Encoders, model, imputers
│       ├── train.csv           # Training data
│       ├── test.py             # Manual test runner
│       └── sample_test1.csv    # Sample test files
├── frontend/
│   ├── src/                    # React + Tailwind frontend
│   │   ├── GameGuardAI.jsx
│   │   ├── App.jsx
│   │   └── index.css, main.jsx
│   └── public/
│   └── vite.config.js
└── README.md

---

📦 Setup Instructions

1. Clone the Repository

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

---

2. Backend Setup (Flask)

cd backend
pip install -r requirements.txt
python app.py

If requirements.txt is not yet created, install manually:

pip install flask flask-cors pandas scikit-learn joblib

---

3. Frontend Setup (React + Vite)

cd ../frontend
npm install
npm run dev

The frontend runs on: http://localhost:5173

Make sure your Flask backend is running on http://localhost:5000

---

📁 Sample Test CSVs

• test_sample_1.csv: Mixed fair and flagged players
• test_sample_3.csv: All clean players
• sample_test_valid_5_rows.csv: 5-row valid structure for testing

---

🧠 ML Model Overview

• Trained on telemetry + system behavior features

• Uses RandomForestClassifier (or similar) for classification

• Handles class imbalance and unseen categorical values

• Prediction includes:

  • Player ID
  • Classification (Cheater / Fair)
  • Confidence Score
  • Behavior Flags

---

📸 Screenshots

Upload + Analyze

Results Dashboard

Demo

31.07.2025_20.34.48_REC.mp4

---

👨‍💻 Authors

• Siddhi Nagapure
• Special thanks to collaborators and dataset providers

---

📄 License

This project is open-source and available under the MIT License.