Rakha X is a full-stack, end-to-end safety ecosystem designed to protect users in high-risk environments.
Unlike standard CRUD applications, this platform integrates real-time audio machine learning, automated emergency response, and AI-driven psychological support into a single cohesive system.
A continuous surveillance module that listens for danger signals.
- YAMNet (TensorFlow): Detects environmental threats like gunshots, glass breaking, and explosions.
- Vosk STT: Analyzes live speech for panic keywords (e.g., “help”, “stop”).
- Automated Logic: Triggers SOS when threat probability crosses a defined threshold.
Automated crisis-response pipeline.
- Twilio Integration: Sends instant WhatsApp alerts to emergency contacts.
- Alert Payload Includes:
- Live location
- Custom emergency message
- Audio evidence clip
AI-powered emotional and psychological support.
- Powered by Gemini Pro
- Provides:
- Emotional first aid
- Crisis de-escalation
- Context-aware safety guidance
Gamified safety education.
- Interactive HTML5 Canvas game
- Teaches identification of:
- Safe zones
- Caution zones
- High-risk areas in urban environments
rakha-x/
├── models/ # Pre-trained TensorFlow / YAMNet models
├── static/ # CSS, JS, game assets, audio files
├── templates/ # HTML (Jinja2 templates)
├── app.py # Main Flask application
├── audio_processor.py # Librosa & Vosk processing logic
├── requirements.txt # Py
└── README.md #