FinSecure AI — Smart Fraud Detection
AI-powered platform that analyses transaction features in real-time, assigns a risk score (0–100) and raises alerts — offline-capable, lightweight, and explainable.
- • Real-time scoring & alerts
- • Lightweight model (RandomForest) with offline mode
- • Simple explainability and risk dashboard
- • Easy to integrate with banks and fintech apps
Fraud Detection — Try It
How It Works
1
Ingest
Collect transaction features (amount, device, location, history) from payment gateway or mobile app.
2
Score
Model predicts fraud probability; thresholding gives fraud flag. Save logs to DB.
3
Act
Trigger OTP, block transaction, or alert analyst via dashboard / SMS / email.
Tech Stack & Deployment
Backend
- • Python Flask API (models/model.joblib)
- • SQLite for logging
- • Joblib for model persistence
Frontend
- • React + Tailwind (this component)
- • Streamlit for quick internal dashboard
- • Vercel to host frontend, Heroku/VPS for backend
Deploy fast: Create a GitHub repo, push React app, and connect to Vercel. Deploy Flask on Render/Heroku or a small VPS. Ensure CORS or proxy is configured.
Tip: For hackathon submit ZIP with frontend/ (this app) and backend/ (Flask + models). Include README.md with run commands.