The use of markerless systems as an alternative for tracking human movement has been growing recently with the promise of allowinf objective analysis outside biomechamical laboratories. However their accuracy might vary from system to system and differ according to the nature of movements and biomechanical indicators. We aim to validate the Apple Augmented-Reality Kit (ARKit) system by comparing the results with a reference optoelectronic system. We investigated the accuracy of ARKit’s measurement of trunk flexion-extension range of motion and the distance between feet during stepping. We carried out laboratory research involving a series of movements, which were captured simultaneously by the ARKit and VICON systems. The results show that from an iPhone positioned parallel to the individual´s frontal plane, the ARKit system can be useful for measuring mediolateral step size but there are significant errors for the trunk flexion extension angle amplitude and anteroposterior step size.