Robotics training data from real human hands at work.
Robots learn manipulation from watching people. We record skilled manual work from the worker’s point of view, with motion sensors, hand pose and step-by-step action labels, in real workshops rather than labs.
Free 10-hour sample first. A person replies within one working day.
Diverse, messy, real manipulation.
Lab teleoperation is clean but narrow. Real workplaces add the variety models need: different tools, materials, lighting, clutter and techniques, done by people who do the job every day.
Hands in view
Head-mounted capture keeps both hands and the work surface in frame for most of each task.
Long-horizon tasks
Whole jobs, not snippets: a garment cut and sewn, a part removed and replaced.
Sensors in sync
IMU and GPS recorded with the video; your own rigs supported if you ship them.
Labels for learning
21-keypoint hand pose per frame, action segments in plain verbs, short captions.
Ready for your training pipeline.
- Video
- MP4 H.264 1080p30, bystanders’ faces blurred
- Sensors
- IMU + GPS sidecar (CSV), time-aligned
- Hand pose
- 21 keypoints per hand per frame, with confidence
- Actions
- Start and end times for each step of the task
- Packaging
- LeRobot or RLDS on request; JSON metadata per clip
- Hardware
- Stereo, depth or multi-camera rigs: you ship, we fit and train
Every minute checked by a person.
ID-checked people
BVN or NIN plus a live selfie matched to the ID, and a bank account in the same name, before any paid task.
Consenting workplaces
Each shop is registered in person by a TrustAm field agent, with the owner’s consent and a location fence.
Live, on-the-job capture
A spoken challenge code, motion sensors and GPS show each clip was filmed live, at work. No camera-roll uploads.
A person reviews every minute
Automatic checks first, then a trained reviewer approves or cuts each minute. Contributors are paid only for approved minutes.
Buyers usually ask.
What robotics training data does TrustAm Data provide?
First-person video of people doing skilled manual work (tailoring, auto repair, phone repair, welding, carpentry, cooking) with IMU and GPS, and optional hand pose, action segment and caption labels, for robot learning and embodied-AI models.
Do you deliver LeRobot or RLDS?
Yes, on request. The default is MP4 plus JSON metadata per clip with a CSV sensor sidecar.
Do you do teleoperation?
Not today. We collect human demonstration video. If you need a specific rig or protocol, send the spec and we will scope it.
How do we start?
Email [email protected] or fill the request form. We record a free sample of about 10 hours to your spec before any paid pilot.
More from TrustAm Data.
Tell us what you need. We send a free sample first.
Email [email protected] with the data type, hours, languages or trades, labels and format. Pricing is quoted per approved hour after scoping.