AI QR PHISHING RISK PREDICTOR

AI-Powered QR Code Phishing Risk Predictor with Multilingual Voice Alert is an intelligent security solution developed for the IBM BOB 2.0 Hackathon. The system helps users identify potentially malicious QR codes before they proceed to the linked destination or payment. The application uses a Streamlit web interface with OpenCV for QR detection and decoding. Users can scan a QR code through a live camera, upload an image, or manually enter a URL or UPI text. The decoded content is analyzed using a RandomForest Machine Learning model combined with 15+ rule-based phishing detection checks. The system classifies every QR code into three categories: SAFE, SUSPICIOUS, or DANGEROUS. It also provides an explainable result by showing the specific reasons behind the detected risk, such as suspicious domains, phishing keywords, raw IP addresses, unusual UPI IDs, or missing payment details. To make security warnings accessible to users across different regions, the system supports 23 languages and can automatically select the language based on the user's location, with a manual language option. Using Google Text-to-Speech (gTTS), the application provides multilingual voice alerts immediately after the QR code is analyzed. By combining real-time QR scanning, Machine Learning, rule-based security analysis, explainable results, location-based language selection, and voice alerts, the project aims to provide users with a simple and accessible way to identify QR-based phishing and payment risks before proceeding.