grand-security

bySeju Chodry

Build a full-stack AI-powered real-time mobile/web security application that detects risky calls, messages, links, and transactions instantly and provides intelligent alerts and automatic actions. 1. Technology Stack (Must Use) Frontend Use Next.js (React) Use Tailwind CSS for UI Use Framer Motion for animations Use Chart.js for risk graphs Use shadcn/ui for UI components Backend Use Node.js with Express.js Use Socket.IO for real-time updates REST APIs for communication Database Use MongoDB to store: User data Risk history Transaction logs Call/message logs AI Microservice Use FastAPI ML Models: Random Forest LSTM NLP: Use Rasa for AI assistant Use scikit-learn for risk scoring Real-time WebSocket based communication 2. Authentication Module Login Page Signup Page OTP Verification Forgot Password Logout Delete Account JWT Authentication 3. Dashboard Risk Meter (Gauge style) Overall Risk Score (0–100%) Risk Status: Safe Low Medium High Real-time updates using WebSockets Charts: Daily risk graph Weekly risk graph Quick buttons: Scan Link Check Number Analyze Message 4. Message Detection Module Monitor incoming messages Detect phishing links NLP scam keyword detection Risk percentage calculation Show reason for risk Auto actions: High risk → Block Medium risk → Ask user Low risk → Notification 5. Call Detection Module Incoming call risk detection Spam number check Real-time popup alert Risk score Auto block high-risk calls User approval for medium risk 6. Link Detection Module URL scanning Domain reputation check AI risk scoring Real-time alert before opening link Risk percentage display 7. Transaction Monitoring Detect online payment activity Set money threshold Alert before payment completion (1–2 sec) Risk scoring for receiver Suspicious transaction detection 8. AI Assistant Chat-based AI assistant Built using Rasa NLP Answer: Why call blocked? Why message risky? Why transaction unsafe? Suggest actions: Block Report Ignore 9. Decision Engine High Risk → Auto block Medium Risk → Alert + ask user Low Risk → Notification only Learning from user behavior 10. Settings Page Enable/Disable: Call monitoring Message monitoring Transaction alerts Link detection Permission manager: SMS Calls Contacts Storage Camera Notification toggles Risk threshold slider 11. Real-time Requirements Detection within 1–2 seconds WebSocket live updates Background monitoring service Push notifications 12. UI Requirements Modern clean UI Dark/Light mode Animated risk meter Graph analytics Smooth transitions 13. AI Model Requirements Random Forest for classification LSTM for sequential message detection NLP keyword detection Risk scoring output: 0–30 → Low 31–70 → Medium 71–100 → High 14. Extra Features Alert history page Manual scan option Risk explanation popup User feedback learning Emergency protection mode Expected Output Full frontend (Next.js) Backend API (Node.js) AI microservice (FastAPI) Database schema (MongoDB) ML model integration Real-time alert system Fully working project