RoboCap Head-worn Rig with 6 Cameras, 2 IMUs for Data Capture

In order to train robots and AI in general, you are going to need high quality data. We have already covered plenty of hand-worn sensors that help you collect data. The RoboCap is a head-worn rig with 6 cameras and 2 IMU sensors. You can use something like this to capture spatial data and prepare exports for Rerun and Hugging Face. It can sync wirelessly to RoboWrist for syncing POV and wrist data.

Thanks to Grounded API, you will be able to enhance your data with wrist & finger tracking, SLAM, and depth estimation. The rig is designed to be comfortable to wear for prolonged sessions. Thanks to this, you will be able to generate rich data to train without having to develop an expensive pipeline.

4/ Purpose-built for RoboCap, Grounded API turns raw sessions into structured data:

> SLAM: 3D camera tracking
> Hand tracking: finger + wrist motion
> Sub‑cm depth at manipulation range
> Per‑session calibration for lab‑level accuracy in the wild

The API also supports easy… pic.twitter.com/DksPKngnDp

— BitRobot 🦾 (@BitRobotNetwork) September 16, 2026

[where to get it]