Acquire · 01
Egocentric Video Capture
Wearable gaze-oriented camera frames (Ray-Ban Meta, chest mounts). A brief can specify suburban households, workshops, or commercial kitchens.
What we do
From a brief you write to a dataset you can load
Four stages take a capture brief and turn it into training data: we acquire it, annotate it, prepare it, and check it. Each stage below says what runs today and what does not. The annotation pipeline runs now; the Capture Partner network is still being built.
Delivery formats
we deliver LeRobot v2.1, HDF5
(EgoDex-style paired files) and the open FH-Ego v1 JSON
schema; Zarr v2 and TFRecord (RLDS-style) are produced on request.
Every format is verified by an automated export bench before we quote it.
We do not offer MCAP or WebDataset — MCAP is a robotics logging
container rather than a training format, and labs shard WebDataset
internally; we would rather tell you which is which.
Data Ingestion & Wearable Capture
We give Capture Partners in India a wearable camera and our capture app, and they record the task being done. The footage comes from roughly where your robot's own camera sits — that is the whole point of capturing it this way.
Acquire · 02
Off-the-Shelf Licensing
Licensing of already-captured, bimanually annotated manipulation clips where a brief has been run. The catalogue lists the real availability status of every capture specification.
Acquire · 03
Synthetic Generation
Generated environments for co-training, covering the edge cases that are hard or unsafe to film. Research phase — not commissionable yet.
Ready to scope your training dataset?
Send the task list, camera perspectives and format requirements. We come back with a scoped brief — what would be captured, at what tier, in what volume, and what it costs. There is no inventory to pick from; every engagement is specified before it is collected.