🛡️ Fraud Eye

Advanced Fraud Detection and Analysis System with Apple-style UI.
Analyse financial transactions in real time, score them with AML heuristics, visualise account networks, and get instant alerts — all in one clean dashboard.

React · Vite · Tailwind CSS · Framer Motion · Node.js · Express · MongoDB · Socket.io

Every transaction is scored 0–100 against 8 Anti-Money Laundering rules:

Risk levels: 🟢 Low (0–29) · 🟡 Medium (30–59) · 🔴 High (60–100)

• Live stat cards: total transactions, high/medium risk counts, reviewed count
• 7-day transaction volume area chart
• High-risk transactions bar chart by day
• Top 5 flagged accounts by alert count and volume
• Recent transactions feed

• Add individual transactions with full AML analysis on submit
• Bulk upload via CSV paste — processes hundreds at once
• Filters: risk level, search by TX ID or account, date range
• Paginated table with inline AML flag breakdown
• One-click review action (analyst/admin only)

• Force-directed SVG graph of medium/high risk transaction relationships
• Nodes = accounts, links = transactions coloured by risk level
• Pan, zoom, hover tooltips showing account stats

• Socket.io pushes toast notifications instantly when a high/medium risk transaction is submitted
• Alerts page: dismiss, filter active/dismissed, colour-coded by severity

• Export filtered transactions to CSV (by risk level or all)
• Preview report before downloading
• Bulk CSV upload with per-row result feedback

• Full dark/light toggle per user account (persisted to database)

• JWT-based login/register
• Three roles: Analyst (can review), Admin, Viewer (read-only)

• Node.js 18+
• MongoDB URI (optional — runs with in-memory DB out of the box)

git clone https://github.com/manny2341/fraud-eye.git
cd fraud-eye

# Install all dependencies
npm install
cd backend && npm install && cd ..
cd frontend && npm install && cd ..

# Start both frontend and backend
npm run dev

• Frontend → http://localhost:5173
• Backend → http://localhost:5001

No MongoDB needed. The app automatically uses an in-memory database if no MONGO_URI is set. Data resets on restart.

To use a real database, add your MongoDB Atlas URI to backend/.env.

Create backend/.env (copy from backend/.env.example):

PORT=5001
MONGO_URI=mongodb+srv://<user>:<password>@cluster.mongodb.net/fraudeye
JWT_SECRET=your_secret_here
JWT_EXPIRES_IN=7d
CLIENT_URL=http://localhost:5173

fraud-eye/
├── backend/
│   └── src/
│       ├── models/          # User, Transaction, Alert
│       ├── controllers/     # Auth, Transaction, Alert logic
│       ├── routes/          # /api/auth, /api/transactions, /api/alerts
│       ├── middleware/       # JWT auth, role check
│       └── utils/           # ApiError, ApiResponse, asyncHandler, AML heuristics
├── frontend/
│   └── src/
│       ├── pages/           # Dashboard, Transactions, Graph, Alerts, Reports
│       ├── components/      # Sidebar, Layout, StatCard, RiskBadge
│       ├── context/         # AuthContext (JWT + theme)
│       ├── hooks/           # useSocket (real-time alerts)
│       └── utils/           # Axios API client
└── package.json             # Runs both with concurrently

_{Built by manny2341}