A living dataset of 54,980 egocentric human episodes totaling 1,213 hours across 5,061 tasks. Continuously expanded by contributors worldwide, validated by 4 research labs and 3 industry partners.

Record everyday tasks — cooking, cleaning, assembly, gardening. The app guides capture quality, validates on-device, and uploads to the EgoVerse with dense annotations.

Contributors earn $HAN for validated egocentric data. Quality scores, task diversity, scene novelty, and annotation density determine reward multipliers. Staking unlocks governance.

Data provenance, contributor identity, and quality scores registered on-chain as verifiable assets. Every data ID, model checkpoint, and skill module is traceable and tradable.

Research partners access curated datasets to train generalizable manipulation policies. Human-to-robot transfer validated across diverse embodiments — bimanual, single-arm, mobile.

Token holders govern reward curves, data quality standards, partnership decisions, and protocol upgrades. Robotic intelligence belongs to all — so does its governance.