Warehouse, packing and retail task data
Picking, packing and shelf stocking follow a procedure and repeat all day. That makes them a natural early job for robots, and a task where models need many objects and many hands.
We recruit people who do this work for a living and film them at their own workplace, on your equipment or ours, with the owner's approval in writing.
Bins, shelves, totes.
By order, size, destination.
Boxing, filler, sealing, labelling.
Pallets, trolleys, vans.
Unpacking, facing, rotation.
Opening, checking, counting.
Send the tasks, objects and edge cases your model is missing.
Real clips with metadata, labels and hand pose, in your format, before you commit.
MP4 and JSON by default. LeRobot, HDF5, RLDS, MCAP or your own schema on request.
On the device, on upload, by a reviewer and against earlier footage. Rejected clips are not billed.
Scripted slips, drops and retries, each one time-stamped.
At the operator's own workplace: stockrooms, packing rooms and shops, with the owner's approval in writing. We recruit in Bali, Mexico, China, Brazil and Denmark.
Yes. Ask for a sample pack: real clips with metadata, labels and hand pose in your format, to load into your pipeline before you commit.
Video with metadata for every clip: task, location, operator ID and consent record, plus a QA report for every batch. On request we add task labels, action timestamps and hand pose data.
We price per accepted hour. The rate depends on the task, the sensors and the rights. Tell us your volume and budget and we scope the project to fit.
Ask for a sample pack, or send the tasks and the spec. We reply with a pilot proposal covering scope, timeline and price.