RakShield™

RakShield™ turns an Android phone into a private, local personal safety companion. It uses phone sensors, local physics-based calculations, and feedback learning to help surface unusual environmental changes based on device sensor data. It is not a guarantee of safety, protection, detection, emergency response, or any specific service. It is not a replacement for common sense, personal judgement, emergency services, professional security, smoke alarms, carbon monoxide alarms, or medical care. Place the phone where you want it to monitor, follow the calibration instructions, and RakShield™ learns the normal baseline for that room and device. While armed, it watches for meaningful changes across available sensors, such as sudden sound, vibration, movement, light change, magnetic-field change, charging change, or nearby device handling. Privacy-first by design: - Sensor data stays on the device. - Raw audio is not uploaded to a backend. - Audio is used only for short-lived local calculations and labels. - No photographs are captured. - No camera access is required for monitoring. - Learning happens locally from feedback and history. What makes RakShield™ different: Many phone alarm apps focus mainly on touch or motion. Many sensor tools show measurements but do not provide a guided safety workflow. Many SOS-style safety apps depend on contacts, location sharing, cloud services, or camera/video recording. RakShield™ is designed around a different loop: local readiness checks, deliberate calibration, multi-sensor monitoring, alert evidence, user feedback, local learning, grace periods, and safer disarm friction. That means RakShield™ does more than make noise after a single movement. It compares signal intensity, duration, steepness, sensor agreement, baseline quality, AI-assisted audio labels, and local feedback history. When it needs attention, it shows a score, severity, graph, and reason timeline so the user can understand why the alert happened. Key features: - Simple Disarmed, Arming, Armed, and Alert states - Device readiness checks before arming - Test Siren action to verify the local alert sound channel - Placement and calibration guidance before monitoring starts - Local sensor fusion for sound, movement, light, magnetic field, charging, and placement context - Alert score, graph, and reason timeline - Feedback options for correct alerts, false alarms, missed concerns, late alerts, and uncertain events - Local learning history so users can see how feedback is improving the app - Pickup and feedback grace periods so phone handling does not immediately cause another alert - Safer disarm flow with deliberate slide and hold confirmation - Support, privacy agreement, developer contact information, and developer app link inside the app RakShield™ starts with built-in sensitivity and false-positive guardrails, then uses local feedback to adjust future decisions on this device. Correct alerts, false alarms, missed concerns, late alerts, uncertain feedback, and quiet sessions can all help tune future thresholds. RakShield™ does not claim to identify intent or detect every environmental threat. False alarms and missed events are possible even when the app is working as designed because results depend on phone model, available sensors, sensor sensitivity, placement, permissions, room conditions, battery settings, operating-system restrictions, environmental variables, and the amount and quality of local feedback data. Sensor issues or unsupported sensors can reduce accuracy or cause malfunction. RakShield™ is a companion to support user awareness and judgement, not a replacement for either. The user remains responsible for safety decisions, emergency decisions, and how they respond to alerts.