Field capture

Training data for the work robots will learn.

Praxome records skilled operators doing real work and ships the result as standardized training data with hand-pose ground truth, for robotics and embodied AI teams.

01 · Datasets

What we capture

SKU: PRX-01

CLOSED BETA

GRIP

Manipulation / VLA Fine-Tuning

First-person RGB-D and 3D hand-pose ground truth from operators performing skilled tasks in real environments, such as lab diagnostics and mineral exploration.

Format: Egocentric RGB-D · 3D Hand Pose
Capture: Single head-mounted stereo depth stream with continuous 3D hand-skeleton reconstruction

PRX-01

CLOSED BETA

SKU: PRX-02

COMING SOON

FRAME

Whole-Body Dynamics & Tactile Contact

Multi-view spatial capture combining head-mounted and ambient room RGB-D streams with full-body kinematic tracking and hand contact pressure metrics.

Format: Multi-view 3D RGB-D · Full-Body Kinematics · Tactile Pressure Arrays
Capture: Dual-perspective vision with continuous joint positioning and surface force mapping

Training data for the work robots will learn.

GRIP

FRAME

Register your interest