A steady hand, at component scale.
Board inspection, fine tooling and intricate repair decisions.
Robots need to learn from real skilled work. The people doing that work deserve a share in the value it creates.
Most large-scale video collection focuses on everyday chores. We are testing a different proposition: robotics data drawn from the complexity of skilled trades, recorded with permission and licensed on fair terms.
No commercial dataset is available yet. We are validating the first pilot.Objects change. Tools differ. Every workshop has its own rhythm. That's the variation we're interested in documenting.
Board inspection, fine tooling and intricate repair decisions.
Material control, sewing, cutting and repeated adjustment.
Motorcycle diagnosis, maintenance and repair in real shops.
Tool use, welding, fitting and precision fabrication.
Workshops document the tasks they already perform. Filming starts only after site permissions, worker consent and recording terms are agreed.
Review footage for bystanders and sensitive information, remove unsafe material, then label tasks and assess quality before it reaches a buyer.
Approved datasets would be licensed under written agreements with a receipt trail, contributor attribution and transparent royalty calculation.
We are exploring a collective that pays an upfront fee for approved hours and shares eligible license revenue with contributors. The calculator shows one possible arrangement, not a commercial offer.
Assumes a 200-hour dataset and equal weighting of approved hours, before fees, refunds and taxes. Rates, revenue splits and licensing terms are not yet agreed.
Recording rights, workplace permission and bystander privacy need proper evidence before any clip is accepted.
Raw recordings should be private. Only reviewed, permitted assets can be included in a commercial dataset.
Blockchain may help audit licensing and settlement later. It cannot prove consent or enforce a license by itself.
Whether you run a workshop or build robots, tell us what you're working on. We're looking for a small number of collaborators to shape the first paid dataset pilot.