Our Air Glove is a glove-shaped controller that captures hand motion in real time — it becomes the user's hands in VR, and a way for robots to learn from human demonstration.
We started in medical XR: VR training for clinicians, a digital therapeutic for mild cognitive impairment (currently in exploratory clinical trials), and simulation content like dispatch helicopter CPR. We're GMP certified for medical device software.
The same hand data that makes VR feel real turns out to be what robots need to learn manipulation. So we're now also building tools around physical AI — hand motion capture, teleoperation, and embodied learning datasets.
Based in Seoul.
A ROS-independent teleoperation stack that turns human hand & arm motion into robot joint commands, plus the assets to drive them.
| Repository | What it is |
|---|---|
| whatslab-sdk | Teleoperation core — receivers, hand/arm retargeting, IK, calibration. Pure Python, no ROS. |
| whatslab-ros2 | Thin ROS 2 (Jazzy) wrapper exposing the SDK as nodes. |
| dexterous-hand-urdf | Curated URDF models for dexterous robot hands. |
New here? Start with whatslab-sdk.