Household
Everyday tasks in your own home
Wiping a counter, folding laundry, loading a dishwasher. The manipulation robots need most and have the least footage of.
- Clip length
- 5–30 s
- Per clip
- $2–$4
Data program 01Rev 04
LEO Robotics pays people to film everyday real world tasks. Robots learn manipulation by watching real work being done, and that footage barely exists. Record a clip from your phone, and once it clears review it settles in tokenized stock on Robinhood Chain.
Pipeline
Nothing decorative and nothing to sign up for. A clip moves from your camera to settlement without an account in between.
No account, no password, no email. Your wallet address is the only identity the system keeps, and it is where settlement lands.
Record straight from the phone in your hand, from your own point of view. Five to thirty seconds, one task per clip.
Every clip is watched by a person against the task brief. Clear framing, one task, no other people in shot.
Approved clips settle from the treasury in tokenized stock on Robinhood Chain. Amounts are priced in USD, paid in stock.
Task lineup
The same brief and the same review across the range. What changes is how hard the task is to film and what it pays.
Everyday tasks in your own home
Wiping a counter, folding laundry, loading a dishwasher. The manipulation robots need most and have the least footage of.
Hands on work with living things
Pet grooming, feeding, gardening, washing a car. Unpredictable subjects, which is exactly what makes the footage valuable.
Tool use with a real outcome
Plumbing installs, patching walls, fluid changes, furniture assembly. Long horizon tasks with tools in frame.
Rates are set per task and per wallet, and every task has a pool that can run out. What you see on the record page is what applies.
Capture standard
One sheet, eight lines. Clips that meet it clear review first time; clips that do not are the only ones that come back.
| 01 | Framing | First person, camera held or worn at eye level |
|---|---|---|
| 02 | Duration | 5 s minimum, 30 s maximum, one task per clip |
| 03 | Resolution | 1080p or better, 30 fps or better |
| 04 | Lighting | Even enough that both hands stay readable |
| 05 | Subjects | Your own space. No other people, no documents in shot |
| 06 | Audio | Optional. It is stripped before the clip enters the set |
| 07 | Format | MP4 or MOV, straight from the phone, no edit |
| 08 | Review | Checked by a person against the task brief |
Questions
The footage you record is what makes the models work, so the people producing it should hold the upside rather than a one time fee. Approved clips settle in tokenized stock, which can fall as well as rise.
Yes. Any standard EVM wallet works. It takes a couple of minutes to set one up and there is nothing to fund, since payment only ever moves toward you.
Approved clips become robot training data. Film only your own space, and keep other people and anything identifying out of frame. You can ask us to delete a clip at any time.
Usually inside a day. Every clip is reviewed by a person, so a busy queue can push it further. You can watch the status of each clip on the record page.
Connect a wallet, film one task, and see it through review. There is nothing to install and nothing to pay.
Start recording