Open a sample.
Explore every layer.

Browse real sample clips from Midcentury datasets, paired with the signals and metadata used for training.

Egocentric Vision

First-person activity video paired with synchronized IMU, task metadata, and temporal annotations. Spanning 20,000+ tasks and dozens of everyday environments.

View samples →

UMI Data

Head and wrist recordings of gripper tasks: RGB and stereo cameras, IMU, fused head and tool-centre-point pose, and gripper state.

View samples →