First-person capture, specified to a brief, for models that have to reason about contact, force and spatial layout rather than read about them.
A capture specification pairs first-person video with the telemetry a policy needs (IMU streams and a 3D camera trajectory), and every clip is annotated against the open FH-Ego v1 schema. Nothing here is stock: the brief defines the task, the environment and the rig before anything is recorded.
Every clip carries a SHA-256 consent receipt in a hash-chained, DPDP-aligned ledger, so a buyer re-verifies one clip instead of trusting a badge. Automated face redaction (YuNet where the host provides it, an OpenCV cascade where it does not) runs server-side at ingestion, before any human review, and fails closed; it refuses to emit a clip rather than emit one un-redacted. Licence plates are not covered. We hold no third-party attestation: no SOC 2, no ISO 27001.
Scoping is a conversation, not a price list: environment, session length, sensor rig and annotation tier each move the number. Capture Partners are paid from a published floor of ₹229/hr of accepted capture, by UPI on QA approval, no agent in the middle and no deduction. Those are the terms partners will be paid on, not a payout rail already running: the Capture Partner network is in pilot build-out and no payment provider is integrated.
Physical AI needs cause, effect, force and spatial layout: material compliance, weight distribution, tool mechanics. None of it survives a text prompt or a web scrape. Below are the streams a capture specification names, and the layers the annotation pipeline derives from them. Model-generated output is derived, never measured, and is labelled that way wherever it appears.
| Signal | What it is |
|---|---|
| Egocentric video | 60fps high-resolution first-person capture, on the rig the brief specifies |
| IMU telemetry | High-frequency accelerometer and gyroscope streams, time-aligned to the video |
| Camera trajectory | 3D camera pose over time (SLAM), so motion is metric rather than relative |
| Hand pose | 21-joint bimanual 3D hand skeletons (MediaPipe) Derived |
| Wrist trajectory | Camera-relative wrist vectors, executable by an actuator Derived |
| Action boundaries | Temporal verb-noun segments marking where one action ends and the next begins Derived |
Tell us the task, the environment and the policy you are training, and we scope a capture brief against it. The catalogue The catalogue lists capture specifications you can commission, with target volumes and their real availability status.
World models need contact and layout, not just pixels. Tell us which physical property your model is failing on and we will scope capture against that, rather than against a volume target.