An operator, not a data broker.
Kinema is a bespoke capture operation, not a reseller of someone else's corpus. It's run by a practitioner who builds and operates volumetric multi-camera rigs, wearable multi-sensor capture, and 4D Gaussian-splatting pipelines.
Why us
The robot-learning field has a data bottleneck: compute is abundant, but action-paired, rights-clean, diverse sensorimotor data is not — and it can't be scraped. We come at it from the capture side. Years of volumetric and 4DGS work mean we can put the right rig on the right task, extract calibrated 3D action and hand-pose tracks, and deliver them in the format your pipeline ingests.
Honest about the stage
This is an early-access pilot cohort, capped at 12. We don't claim a standing inventory of petabytes — we claim a real capture capability and bespoke, made-to-order collection. Every figure on this site cites a public source or is labeled a pilot estimate. Early partners help shape the capture protocol and get exclusivity options.
Talk to the person who captures the data
No SDR funnel. Apply to the cohort and you'll hear back from a real person within one business day to scope a bespoke capture. Prefer email? pilot@flyreelstudio.eu.