Collector role
Physical AI Data Collector Jobs
TrueLabel accepts collectors for physical AI data collector opportunities where eligible applicants record approved task footage for physical AI and robotics teams.
Overview
Physical AI data collection captures approved everyday task footage that robotics and physical-AI teams use for training. The emphasis is on realistic, multi-step routines completed start to finish so each capture shows a full, repeatable sequence. You record on a phone or approved camera, keep the action continuous, and submit raw clips. Payment applies only to accepted footage.
Collectors should be ready to share location, device model, mounts, language, recording space, and availability. Use a recent smartphone or approved wearable/action camera at 1080p/30fps minimum and keep the camera positioned to follow the whole routine continuously. TrueLabel uses those details to match applicants to role-specific and location-specific collector opportunities. Learn more about physical AI data collector in Mexico, physical AI data marketplace, privacy and consent for video capture.
Physical AI Data Collector job answers
Collector opportunity details
- Role
- Physical AI Data Collector
- Work type
- Remote, opportunity-based multi-step routine capture (independent contractor)
- Common regions
- the United States, Mexico, Canada, and LatAm
- Typical equipment
- recent smartphone or approved wearable/action camera
- Capture spec
- Use a recent smartphone or approved wearable/action camera at 1080p/30fps minimum and keep the camera positioned to follow the whole routine continuously.
- Pay
- $18-$24 per approved hour of usable footage
- Payment basis
- $18-$24 per approved hour of usable footage, paid only for footage the TrueLabel collector QA team accepts on review (usually within 2 business days of upload)
- Review
- The TrueLabel collector QA team, usually within 2 business days of upload
- Last updated
- June 5, 2026
What this opportunity involves
What a physical AI data collector records
Physical AI data collection captures approved everyday task footage that robotics and physical-AI teams use for training. The emphasis is on realistic, multi-step routines completed start to finish so each capture shows a full, repeatable sequence. You record on a phone or approved camera, keep the action continuous, and submit raw clips. Payment applies only to accepted footage.
Core responsibilities for physical AI data collectors
This role is defined by a specific set of capture habits: capture full multi-step routines from start to a clear finished state, keep object handling continuous so each step flows into the next, record realistic, repeatable sequences rather than staged one-off shots, and preserve raw footage so the full routine is available for review. Each is checked during review, so practising them before you submit keeps your acceptance rate high.
What gets accepted versus reshot
Footage is accepted when the routine runs from start to finished state without skipped steps, object handling stays continuous and clearly visible across steps, and footage is raw and represents a realistic, repeatable task. It is sent back or rejected when the routine is cut short or skips steps in the sequence, object handling drifts out of frame between steps, and staged or unrealistic motions that would not transfer to a robot.
How TrueLabel matches physical AI data collectors
For a physical AI data collector, the setup that matters most is concrete: use a recent smartphone or approved wearable/action camera at 1080p/30fps minimum and keep the camera positioned to follow the whole routine continuously. A passing sample proves you can capture a continuous multi-step routine to a clear finished state with handling visible throughout. Your profile should also list location, language, available mounts, recording environment, and weekly availability so TrueLabel can match you to eligible work.
What makes a submission review-ready
The single most common rejection is a routine that is cut short before reaching a clear finished state. Beyond that single failure, a review-ready physical AI data collector clip keeps the task visible from start to finish, follows the brief, avoids private information, and arrives as a raw upload. Test your framing on a short clip before recording the real take.
Matching opportunity types
TrueLabel uses collector profile signals such as location, device, language, capture setup, and sample quality to match applicants with eligible collector opportunities.
| Opportunity | Collector work |
|---|---|
| Capture responsibility | Capture full multi-step routines from start to a clear finished state |
| Capture responsibility | Keep object handling continuous so each step flows into the next |
| Capture responsibility | Record realistic, repeatable sequences rather than staged one-off shots |
| Capture responsibility | Preserve raw footage so the full routine is available for review |
Requirements and review
| Area | What to expect |
|---|---|
| Device | Use a recent smartphone or approved wearable/action camera at 1080p/30fps minimum and keep the camera positioned to follow the whole routine continuously. |
| Accepted when | the routine runs from start to finished state without skipped steps, object handling stays continuous and clearly visible across steps, and footage is raw and represents a realistic, repeatable task. |
| Rejected when | the routine is cut short or skips steps in the sequence, object handling drifts out of frame between steps, and staged or unrealistic motions that would not transfer to a robot. |
| Submission | Raw physical AI data collector files uploaded through the approved TrueLabel collector flow; the TrueLabel collector QA team reviews each one usually within 2 business days of upload and only accepted footage is paid ($18-$24 per approved hour of usable footage). |
Privacy and quality expectations
For this collector role across the United States, Mexico, Canada, and LatAm, good collector work is useful because the recording is clear, complete, and safe to review. Keep the task visible, avoid private information, submit raw files, and follow the opportunity brief before recording. If a project asks for first-person or smartphone video, assume that faces, IDs, payment cards, screens, addresses, private documents, and bystanders should stay out of frame unless the brief explicitly says otherwise.
For additional background, TrueLabel links to public references on privacy and responsible AI data practices. The opportunity brief, collector agreement, and TrueLabel review outcome remain the source of truth for what is accepted, rejected, or paid.
Related collector opportunities
The related opportunities below show how specific collector work is scoped across the United States, Mexico, Canada, and LatAm when TrueLabel has matching work categories.
Related collector opportunities
FAQ
What makes a physical AI data collector submission pass review?
A passing sample proves you can capture a continuous multi-step routine to a clear finished state with handling visible throughout.
What is the most common reason physical AI data collector footage is rejected?
The single most common rejection is a routine that is cut short before reaching a clear finished state. Most reshoots for this role come back to that single issue, so check it on a short test clip before recording the full task.
Are rejected physical AI data collector uploads paid?
For this role, footage is sent back when the routine is cut short or skips steps in the sequence. Payment applies only to accepted work that passes review; duplicate, unsafe, private, edited, or off-brief submissions are not eligible.
Apply for physical AI data collector opportunities
Join the TrueLabel collector network to be considered for physical AI data collector, smartphone video, wearable camera, and hand-object interaction opportunities.