A next-gen, voice-secured digital wallet built to protect users in high-risk scenarios β with real-time AI voice recognition and a duress-safe mode.
SmartPay is a full-featured digital wallet application designed for both convenience and safety. It introduces a game-changing innovation: Voice-Activated Safe Mode (VASM) β which protects users under threat by instantly hiding sensitive financial data, locking transactions, and optionally sending silent alerts.
This project demonstrates a deep understanding of secure financial systems, modern front-end architecture, and AI-powered user safety.
In high-crime environments, users may be forced to open their mobile wallets under threat. If a criminal sees the real balance or is able to initiate transfers, the userβs safety is at risk.
Most digital wallets don't offer any safety fallback under coercion β SmartPay does.
- ποΈ Voice-activated panic mode with offline voice recognition
- π Fake balance mode triggered under duress
- π Transaction lock during panic mode
- π© Silent emergency alerts to trusted contacts
- π Dual PIN system (primary for login, secondary for exiting safe mode)
- π° Secure wallet management (add, send, receive funds)
- π Transaction history with filters and sorting
- π₯ User profile and KYC integration
- π PIN and biometric authentication
- π§Ύ Bill payments (e.g., airtime, utilities)
- π Multi-currency support
- π² Responsive UI for mobile and desktop use
- User logs in via PIN or biometric
- Dashboard loads:
- Real wallet balance
- Quick actions: Send, Receive, Pay Bills
- Recent transaction list
- Settings and profile access
- User says a predefined panic phrase (e.g., "Please donβt hurt me")
- Voice recognition module detects the phrase
- App enters Safe Mode:
- Fake balance is displayed (e.g., R0.00)
- All transfers and payments are disabled
- UI mimics normal functionality (decoy mode)
- Optional silent alert is sent to a predefined contact or admin
- To exit Safe Mode:
- Logout required
- Re-login with secondary secure PIN
| Layer | Technology Stack |
|---|---|
| Frontend | Angular 19, TypeScript, CSS, Bootstrap 5 |
| Voice Capture | Web Audio API, Microphone Access |
| Voice AI | Vosk (Offline), TensorFlow.js |
| Backend | ASP.NET Core Web API (C#) |
| Authentication | PIN + JWT |
| Database | SQL Server |
- Offline voice recognition using Vosk
- Fast local processing for panic phrase detection
- Designed to work even in low-network or offline conditions
- Future scope: vocal emotion detection for smarter triggers
- Real-world problem solving: Addresses safety in financial technology
- End-to-end thinking: Full-stack design with security, UX, and AI
- Scalable architecture: Modular Angular components, secure .NET API
- User-first design: Focus on usability, accessibility, and privacy
π₯ Demo walkthrough and screenshots of the wallet UI and Safe Mode in action.
- β Backend APIs completed and secured (ASP.NET Core)
- β Angular 19 UI development underway (modular, responsive components)
- π Voice recognition model integration in progress (using Vosk / TensorFlow.js)
- π Mobile responsiveness testing and optimization (Bootstrap 5)
- π Demo video recording and live deployment planned
I'm open to:
- Collaborations with other developers
Feel free to fork the repo, open an issue, or send a message.