đ¯ Mission
VocalGuard is an AI-powered solution designed to detect and prevent vishing (voice phishing) attacks in real-time. Our mission is to protect individuals and organizations from sophisticated phone scams that exploit emotional manipulation and urgency tactics.
đŦ Technology
VocalGuard employs a multi-faceted analysis approach:
- LTS (Linguistic Threat Score): Analyzes transcript content for scam-related keywords and patterns using advanced NLP models
- EMS (Emotional Manipulation Score): Detects emotional pressure tactics through voice energy, pitch variation, and speaking patterns
- VAS (Voice Authenticity Score): Identifies synthetic or manipulated audio using spectral analysis and harmonic ratios
The combined risk score provides a comprehensive assessment: Safe (<0.4), Suspicious (0.4-0.7), or Risky (>0.7).
đĨ Team & Project
Team Name: The Newtonians
Team ID: 11601
Chapter: Global
Challenge Category: Cyber-Technology & Security
Team Members:
- Tawhid Bin Omar
- Tapantor Das
Institution: St. Joseph Higher Secondary School
VocalGuard was developed as part of the Conrad Challenge, focusing on innovative solutions for cybersecurity and consumer protection. Our interdisciplinary team combines expertise in machine learning, audio processing, and user experience design to combat the growing threat of vishing attacks.
đ Features
- Real-time audio analysis and risk detection
- Multi-modal AI assessment (linguistic + acoustic + voice authenticity)
- User-friendly dashboard with detailed analytics
- Secure cloud-based processing with Firebase
- Usage tracking and threat prevention statistics
đ Privacy & Security
We take your privacy seriously. All audio files are processed securely and temporarily. No recordings are permanently stored without your consent. User data is encrypted and protected by Firebase security rules.
đ§ Contact
For questions, feedback, or partnership inquiries, please reach out to us: