One toolkit for egocentric video across any device

Egocentric video collection for robotics, activity recognition, and embodied AI is hampered by incompatible SDKs, camera access policies, and proprietary platforms across different devices. EgoKit provides a single toolkit that exposes the same recording workflow on six heterogeneous host devices (Android phones, iPhones, iPads, smart glasses, XR headsets) and produces locally stored video in a uniform log format. On XR headsets, it captures synchronized head pose and 26-joint hand tracking aligned to video. Companion accessories—two wrist cameras with mounts, head strap, and USB-C hub—add multi-view capability without custom fabrication. Code and design specifications available at egokit.chuange.org.