PLATFORM

Dual-stream capture and an end-to-end pipeline.

Our platform combines proprietary omnidirectional capture with a complete processing pipeline that transforms real-world footage into VLA and VLM-ready training datasets.

The data layer Physical AI is missing.

Synchronized dual perspective

Egocentric and 360° exocentric streams in a single pass. Every demonstration captured from inside the actor and from the room — simultaneously.

Real depth, not estimated

Depth and geometry come from the sensor, not from a model's guess. 7mm accuracy at 5 meters, 0–20m range, captured live. What ships into USD is what we measured.

Real → USD → Real, fully closed

Real-world capture exports to OpenUSD with physics, runs in any USD-compatible simulator, validates back against real-world deployment.

Synchronized dual-stream capture, in a single pass.

Alia 360° omnidirectional camera plus GoPro egocentric capture — 32+ patents, 8 years of production deployment, peer-reviewed at CVPR.

Alia Specs

Full 360° coverage with long-range, high-resolution 3D depth in a single compact unit. No stitching, no blind spots, no multi-sensor calibration.
360°
Horizontal FOV
6912×3072
Stereo Resolution
120m
Detection Range
30fps
Frame Rate
Image & Depth
Vertical FOV170°
Depth Range0cm – 20m (no blind spot)
Depth Accuracy7mm at 5 meters
OutputReal-time RGB + Depth
Edge AI & Durability
ProcessingEmbedded edge AI
Human Detection120m range
Facial Recognition40m range
Ingress RatingIP67
CompatibilityUbuntu, ROS, OpenCV

Purpose-Built for Physical AI Data

Developed from breakthrough research in computational imaging at IIIT Hyderabad and refined over eight years of production deployment, the synchronized dual-stream system captures the complete spatial context that humanoid robots need.

🎥

Dual-Stream Capture

Synchronized 360° exocentric (Alia) + egocentric (GoPro) in a single pass. Complete spatial context — what the robot sees and how it appears in the scene.

🔬

CVPR-Pedigree Optics

Single-shot 360° stereoscopic design from IIIT Hyderabad, peer-reviewed at CVPR 2016. The optical breakthrough behind every dataset we ship.

🏗️

8 Years in Production

Deployed across AMRs, surveillance, and smart city. Real-world failure modes already engineered out before the first humanoid dataset.

🛡️

32+ Patents

A decade of protected optical IP. The capture platform is the moat — every downstream product inherits its quality.

From real-world capture to training-ready data.

Our end-to-end platform transforms real-world captures into VLA and VLM-ready training datasets — delivered in USD, LeRobot/RLDS, and Open X-Embodiment formats.

01

Dual-Stream Capture

Synchronized Alia 360° exocentric + GoPro egocentric cameras with full RGB + depth in real environments

02

Multimodal Annotation

AI-assisted (SAM2, Grounding DINO) + human QA delivers vision, language, and action labels — 10× faster than traditional 3D annotation

03

3D Reconstruction

3D Gaussian Splatting creates photorealistic scenes with all annotations preserved — ready for simulation conversion

04

Simulation Conversion

Automated USD export with physics properties for any USD-compatible simulator

05

Synthetic Generation

1,000+ frames/hour with domain randomization — all modalities and skill transfer demos preserved

06

Real-World Validation

Continuous verification: sim-to-real transfer rates, manipulation success, and skill transfer effectiveness

Vision Data

  • Synchronized ego + 360° exo video with depth
  • Object segmentation with instance IDs
  • 6DOF object poses
  • Manipulation affordances (grip types, approach vectors)
  • Human and robot demonstrations in 360° view

Language Data

  • QA pairs describing objects, actions, and scene elements
  • Action summaries for every sequence
  • Spatial relations between objects and actors
  • Context descriptions for scene understanding
  • Ready for VLA instruction following

Action Data

  • Full trajectories for every actor in 360° scene
  • Movement paths with timestamps
  • Interaction sequences showing manipulation
  • Demonstration labels for skill transfer
  • Kinematic data where available
🟢 OpenUSD Native, physics-ready
🤗 Hugging Face LeRobot RLDS
📦 Open X-Embodiment

Ready to build with real-world data?

Talk to Us →