Consent before capture.
Every person who appears in or creates our data has agreed to it, in writing, before anything is recorded.
- Contributors sign a release in their own language, alongside English.
- The release explains plainly that the data is used to train AI systems.
- Property owners give permission before we film in homes or private workplaces.
- Recognizable people in footage need their own signed release.
- No minors. We do not collect data from or about children.
- No covert recording. Contributors know when they are being recorded.
Privacy by default.
We collect what the spec needs and nothing more.
- Bystander faces in street and field footage are blurred before delivery.
- Contributor metadata uses ranges, such as an age band, rather than exact personal details.
- Names and contact details are kept separate from the delivered data.
- Files are stored with access limited to the people working on the project.
- Sensitive categories, like health or religion, are only collected when a project specifically calls for it and the contributor explicitly agrees.
Fair to contributors.
The people behind the data deserve to be treated well.
- Pay is agreed before work starts and set fairly by local standards.
- Contributors are paid for every file that passes review, with payment records kept.
- Every task comes with a clear written brief and a short trial first.
- Contributors can decline any task, and we never ask anyone to do something unsafe.
Documented for buyers.
Your legal team will ask where the data came from. We make that easy to answer.
- Each batch ships with documentation: what was collected, where, when, and under which consent version.
- Consent records are available to you on request.
- Custom collections are made for one client and are not resold unless agreed in writing.
- Consent and releases follow the data protection law where the data is collected, such as GDPR, CCPA, or Indonesia's UU PDP.
- Read our guide: Consent for AI training data: what to ask your vendor.
Clean rights for archives.
When we license an existing library, we check the rights first.
- We only license material the owner has the right to license for AI training.
- Every archive goes through a rights review before it is offered to buyers.
- Owners keep ownership. Licenses define exactly how the material can be used.
Questions about how we collect?
We're happy to walk your team through our consent process and documentation.
Get in touch