This release makes KinesteX smarter and more reliable where it matters most: more accurate calorie tracking, a more powerful AI model system for motion analysis, and a smoother Timed Up and Go (TUG) assessment experience on mobile.

Calorie Tracking Accuracy

Calorie calculations now use MET (Metabolic Equivalent of Task) values that are tightly linked to each exercise sequence. This delivers more accurate and scientifically grounded calorie estimates for both rep-based and timer-based exercises.

We also separate active exercise time from idle time (user in frame but not exercising). Idle time is tracked at a baseline metabolic rate, so total calories now better reflect actual effort instead of just session duration.

Enhanced AI Model System

The new AI model system is now fully supported and rolling out across more exercises. This architecture enables:

• Faster delivery and updates of exercise models

• Support for enhanced configurations with custom landmark and angle setups

As more exercises migrate to this system, you can expect more precise motion analysis and more consistent feedback across supported content.

TUG Assessment — Better on Mobile

The Timed Up and Go (TUG) assessment has been refined for mobile portrait viewports:

• Start and target zones automatically widen on narrow screens for more forgiving alignment

• Landmark detection now uses percentage-based validation (for example, 5 out of 7 landmarks in a zone) instead of requiring a perfect match

These changes significantly reduce false negatives on phones and make TUG assessments more reliable in real-world mobile use.