Zen-O: Collecting human training data

Zen-O is collecting human data to train physical and embodied AI models.

The project accepts a range of data types, from egocentric video to structured training datasets intended for model supervision and evaluation.

How to participate:

Sign up on the Zen-O web dashboard, contribute real-world labeled data, help train AI models, and receive compensation or rewards tied to contributions.

The initiative positions these datasets to improve physical AI capabilities by leveraging diverse human-collected data.