VoiceShield — Real-Time AI Scam Call Defender

VoiceShield is an AI-powered real-time scam call defender, protecting vulnerable individuals from billions in losses. Traditional caller ID is insufficient; VoiceShield analyzes conversations to detect evolving scam tactics like psychological manipulation and social engineering. This SecOps platform intercepts and analyzes live audio, transcribing it and feeding it to an AI reasoning engine. Instead of keyword matching, it understands context, identifying urgency, untraceable payments, technical exploitation, and authority spoofing. If a scam is detected, VoiceShield issues high-severity alerts, risk scores, and actionable advice (e.g., "DO NOT PAY. HANG UP IMMEDIATELY"). All threats are logged for reporting. Built with React/Vite, Tailwind CSS, Python/FastAPI, PostgreSQL, AssemblyAI, and Google Gemini, VoiceShield boasts a decoupled, cloud-native architecture. Challenges included browser audio handling, deployment, and CORS configuration. We're proud of bridging unstructured audio with actionable intelligence, creating an intuitive UX/UI, and the AI's ability to identify "zero-day" scams. Learnings encompass the MediaRecorder API, asynchronous Python, relational databases, and advanced prompt engineering for structured JSON output. Future plans include mobile OS integration, carrier-level deployment, and multi-language support to end phone scams.