ScamurAI

Inspiration

With the growing rise of online payment fraud, we wanted to create a solution that goes beyond traditional fraud detection. Most systems are centralized and reactive. We were inspired to build a system that is proactive, collaborative, and powered by AI—something that could adapt and scale as fast as the scams do.

What it does

ScamurAI is a decentralized, AI-powered fraud detection firewall. It uses a trained machine learning model to identify suspicious payment behavior in real time. Once detected, the system broadcasts a fraud alert to all connected peers using a Gun.js peer-to-peer network, enabling collaborative fraud defense across all nodes.

How we built it

Backend: Node.js, Express, Stripe API for payment monitoring.
AI Model: Trained a fraud detection model using Python (XGBoost / Isolation Forest).
Peer Network: Built using Gun.js with SEA for identity and signature verification.
Real-time Communication: Fraud alerts are signed and shared via Gun’s decentralized graph.
Dashboard: React.js interface with live fraud monitoring and peer connections.

Challenges we ran into

Managing consistent peer-to-peer communication across different devices.
Ensuring secure and verifiable fraud alert signatures using Gun SEA.
Integrating AI predictions smoothly into a real-time Node.js environment.
Capturing accurate IP and location data in test mode for meaningful geolocation results.

Accomplishments that we're proud of

Successfully deployed a working fraud detection loop from transaction → ML → alert.
Built a decentralized system where every peer contributes to the defense.
Enabled secure identity management and signature verification using Gun SEA.
Created a clean, animated React dashboard to visualize fraud alerts in real time.