
The capture layer for Physical AI.
Abundance is the complete platform for collecting first-person, multimodal demonstration data. Our hardware and software work together as a single pipeline, capturing what your experts see and delivering it as training-ready data.
Request a unit

Introducing the A3 and A4 capture wearables
Together they capture each demonstration in full and in sync — light enough for a full shift and precise enough to train on.
Light enough for
an 8-hour shift.
No straps, no wire — the lightest config in the line. Light enough to forget.



Our platform connects it all
Hardware, a contributor app, and the APS operations platform — one system that carries a session from the moment of capture to training-ready data.




Raw capture in,
training data out.
A vision model reviews every chunk - lighting, hands, productive work. Humans audit these flags.
Compare the line.
| Feature | A3Mono + IMU | A4Stereo + IMU |
|---|---|---|
| Sensor | Mono RGB · rolling shutter | Stereo RGB · global shutter |
| Resolution | 1920×1080 (2MP) | 1920×1200 (2MP) |
| Frame rate | 30fps · constant | 30fps · constant |
| DFOV | 150° | 150° |
| Distortion | 10% | 14% |
| Color depth | 8-bit | 8-bit |
| Codec | H.264 · IPPP | H.264 · IPPP |
| Bitrate | 4–20 Mbps | 4–20 Mbps |
| IMU | 240Hz · 6-DoF | 300Hz · 6-DoF |
| Sensor sync | <1ms · IMU–video | <1ms · all sensors |
| Hand tracking | ~40mm | ~20mm |
| Tactile gloves | No | Per-fingertip force · 12-bit |
| Microphone | USB Audio Class · 48kHz | USB Audio Class · 48kHz |
| Connectivity | WiFi 5GHz P2P + USB-C | WiFi 5GHz P2P + USB-C |
| Storage | External SD or offload through phone client | External SD or offload through phone client |
| Weight | 33g | 56g (130g with board) |
| Dimensions | 85 × 40 × 38 mm | 105 × 56 × 29 mm |
Proven in the field.
Not a prototype. Abundance devices are capturing real demonstrations across the world right now.
- 15,000+
- devices deployedIn the field today
- 7+
- countriesAnd counting
- 33g
- lightest configNo straps, no wire
- 300 Hz
- IMU samplingVIO-grade motion