Comparison

Public egocentric dataset licenses, compared

Which public egocentric datasets you can legally train a commercial model on.

Short answer

Public egocentric datasets vary from permissive to strictly non-commercial. Ego4D permits commercial use under its own agreement [VERIFY], while CC BY-NC and CC BY-NC-ND datasets prohibit it. Firsthand ships a perpetual, buyer-owned commercial license with documented consent.

License matrix

What each dataset actually allows.

Plain-English summary. Always read the current license text before you rely on it — terms change.

DatasetLicenseCommercial useRedistributionNotes
Ego4DEgo4D Dataset License Agreement [VERIFY]Permitted per terms [VERIFY]Restricted [VERIFY]Fitness, not legality, is the usual blocker
EgoDexCC BY-NC-ND 4.0 [VERIFY]NoNo derivativesStrong hand pose, but rules out commercial training
EgoScalePer release [VERIFY][VERIFY][VERIFY]Confirm terms before commercial use
EPIC-KitchensCC BY-NC 4.0 [VERIFY]NoAttribution, non-commercialPopular kitchen benchmark; non-commercial
FirsthandBuyer-owned commercialYesBuyer-owned; derived models yoursPerpetual, irrevocable, consent artefacts included

FAQ

Common questions.

Does CC BY-NC-ND allow commercial AI training?

No. CC BY-NC-ND prohibits both commercial use and derivative works, which a trained model arguably is. Use commercially-licensed data for any product.

Is a trained model a "derivative" of the dataset?

It is a contested legal question and varies by jurisdiction and license [VERIFY]. The safe path for a commercial product is data licensed explicitly for commercial training, like Firsthand.

What license does Firsthand grant?

A perpetual, irrevocable, buyer-owned commercial license, delivered with the batch consent artefacts and a data card recording jurisdictions of capture. See /provenance.

Judge it on your own skill spec.

Get the free 40-episode sample pack, or send us your target skills and hours for a scoped quote.