Skip to main content

How Motra's AI Learns Your Movement Patterns

Motra AI learns your movements over 5 workouts using Neural Kinetic Profiling. Correct errors and record videos to improve accuracy.

Updated this week

Motra uses Neural Kinetic Profiling™ (NKP) to automatically detect exercises, count reps, and track workout metrics from your Apple Watch's motion sensors.

Motra's model improves accuracy by learning from your workout history. After approximately 5 logged workouts, accuracy improves by up to 20%. This learning period is called the Learning Phase.


How the Learning Phase Works

During the Learning Phase, Motra may experience minor errors in rep counting and exercise detection. These errors decrease as the model adapts to your movements. To help improve accuracy during this period, correct any rep count or exercise errors when they occur and work out consistently using Motra.

Factors That May Affect Auto-Detection Accuracy

Auto-detection accuracy can be affected by improper exercise form, very rapid or jerky movements, Apple Watch sensor calibration issues, weak Bluetooth connectivity between your Apple Watch and iPhone, and incorrect wrist or crown orientation. To improve accuracy, make sure your Apple Watch is worn snugly, the app is up to date, and your Digital Crown orientation matches your actual wrist setup.

To reset AI learning and restart the Learning Phase, on your iPhone:

  1. Open Motra and go to Settings > Workout Settings > Auto Detection Control.

  2. Tap "Reset Detection Cache."

    Note: Resetting clears all learned movement patterns. Motra will need approximately 5 workouts to re-learn your patterns.

Did this answer your question?