╔═══════════════════════════════════════════════╗
║ SPECTRA_X1 // PROPRIETARY ACQUISITION RIG ║
║ STATUS: DEPLOYED ║
╚═══════════════════════════════════════════════╝
4K HDR POV CAPTURE FOR EMBODIED AI
SPECTRA-X1 is Robot Flow Labs' proprietary eyewear capture system. It records the world exactly as humans interact with it — providing the foundational input layer for embodied AI training.
To train human-like perception, you need human-grade input.
| PARAMETER | VALUE |
|---|---|
| Maturity | PRODUCTION |
| Version | X1 |
| Deployment | ACTIVE |
| Focus | Dense POV data capture at industrial scale |
┌─────────────────────────────────────────┐
│ VIDEO ........... 4K HDR @ 60FPS DUAL │
│ IMU ............. 6-AXIS SYNCHRONIZED │
│ TRACKING ........ GAZE + DEPTH MAPPING │
│ FRAME ........... LIGHTWEIGHT INDUSTRIAL│
│ STREAM .......... MULTI-SENSOR FUSION │
│ OUTPUT .......... RAW + CALIBRATED │
└─────────────────────────────────────────┘
- DUAL STREAM CAPTURE — Synchronized stereo 4K HDR video at 60fps
- 6-AXIS IMU — Real-time acceleration, gyroscope, and magnetometer data
- GAZE TRACKING — Eye movement and fixation point recording
- DEPTH MAPPING — Spatial context from structured light
- INDUSTRIAL FRAME — Lightweight, comfortable for extended capture sessions
Captured streams feed directly into the Robot Flow Labs data pipeline:
SPECTRA_X1 → RAW_CAPTURE → CALIBRATION → DATA_PIPELINE → TRAINING_READY
spectra-x1/
├── docs/ # Technical documentation
├── calibration/ # Calibration procedures and tools
├── firmware/ # Device firmware (restricted)
└── specs/ # Hardware specifications
- X1 prototype deployed
- 4K HDR dual stream validated
- IMU synchronization calibrated
- X2 design phase (enhanced FOV)
- Multi-unit fleet synchronization
See CONTRIBUTING.md
@misc{robotflowlabs2026spectrax1,
title={SPECTRA-X1: Proprietary POV Acquisition Rig for Embodied AI},
author={Robot Flow Labs},
year={2026},
url={https://robotflowlabs.com}
}Proprietary — Robot Flow Labs. See LICENSE for details.
ROBOT FLOW LABS // SPECTRA_X1 // DEPLOYED
DATA FIRST. MODELS NEXT.