Can a safety app understand the difference between danger and a morning jog?
We’re building Garly, a privacy-first AI safety companion, and one of our biggest challenges is avoiding false alarms.
A phone’s sensors can detect movement, but real life is complicated. Running, travelling by bus, dropping a bag or making a sudden movement can look similar without enough context.
Feedback from our testers led us to develop adjustable sensitivity, Jogging Mode, Bag Mode and configurable stillness detection. We’re now combining accelerometer, gyroscope, impact and movement interruption signals to make protection more context-aware.
The goal is not to replace emergency services or pretend that a phone can understand every situation. We want to create an additional safety layer that users can trust and control.
What would make you trust an automatic protection feature: a cancellation countdown, a check-in notification, conservative detection, or something else?
You can try the live demo here:
Replies