Library

What we deliver.

Samples from real capture programs, grouped by what you get. Every clip on this page was recorded in a working environment, and every asset was built from one. Nothing is staged, and nothing is generated. The exocentric stream throughout is Alia. Customer names are withheld by agreement.

( 01 )

360° exo views with depth

World model data

The full sphere at 16K with metric depth for every pixel, from one camera position. Each sample below is a color panorama and its depth map, from the same instant. The same camera as the datasets below, delivered as a standalone panorama rather than paired with a worker’s view.

Capture
Alia, single shot
Color
equirectangular
Depth
Metric depth at every pixel
Format
Equirectangular video or stills, with the depth map. Agreed per program.

A sample of the panorama library. Full set available under NDA.

( 02 )

VLM datasets

Robot training data

Multi-view video paired with questions and answers about what is happening in the scene. Built to test spatial reasoning, physical reasoning, and understanding of embodied action.

Capture
Ego + exo, synchronized
Annotation
Capability-probed Q&A pairs
Format
Video with paired Q&A. Schema agreed per program.

A sample of the VLM library. The PRISM-100K open subset is on Hugging Face.

( 04 )

Loco-manipulation datasets

Robot training data

Work that involves moving through the space, not just working at a fixed station. Approach, reach, whole-body positioning, and two-handed work at close range — captured from the worker’s point of view and from the room at the same time. Annotated for hand pose, contact, action boundaries, and task breakdown.

Capture
Ego + exo + wrist, synchronized
Annotation
Hand pose, contact, action boundaries
Format
LeRobot (RLDS), Open X-Embodiment
Stocking and replenishment
Grocery · ego + exo + wrist
Pick and pack
Retail store · ego + exo + wrist
Food prep, knife work
Commercial kitchen · ego + exo + wrist

A sample of the loco-manipulation library. Full dataset available under NDA.

( 05 )

Simulation environments

In development

Whole venues reconstructed from Alia capture and rebuilt as simulation scenes, with collision geometry and material properties. Layouts, shelving, fixtures, and lighting come from a real place.

Source
Reconstructed from Alia capture
Physics
Collision geometry, material properties
Format
OpenUSD, 3D Gaussian Splats, or meshes
In development.

We are building this from the same capture that produces the datasets above. Talk to us about the venue you need and we will tell you where we are.

On formats

Where a standard exists we deliver to it: LeRobot (RLDS) and Open X-Embodiment for loco-manipulation, OpenUSD for simulation. VLM, navigation, and panorama deliveries have no single standard, so we agree the schema with you before capture starts.

Tell us what your model needs to learn.

Capture programs, research collaboration, and dataset partnerships.

Talk to us