Recording the world for the machines that will work in it.
Rayform exists because robots are about to leave the lab, and almost nothing about the places they are going has been recorded in a form they can learn from.
What we do.
Rayform builds and licenses a dataset of real environments captured in three dimensions over time. Not scans of empty rooms, but recordings of places behaving: people and robots moving through them, objects being handled, doors opening, light changing.
We capture in the field with calibrated multi-sensor rigs, register every stream into one clock and one coordinate frame, and label every capture with the same vocabulary. The result is one corpus that a team can query by place, task, object, or condition and stream straight into training.
We are in private preview. The teams we work with now shape what we capture next.
Where we capture.
Variety is planned, not accidental. Each kind of place is captured across conditions, so a model learns the world and not one room.
- Homes
- Kitchens, hallways, living rooms: the cluttered, lived-in spaces where household robots will work.
- Factories
- Assembly lines, loading bays, and the narrow paths between machines.
- Mines
- Dust, low light, uneven ground, and moving equipment, above and below the surface.
- Open terrain
- Fields, construction sites, and the outdoors, for navigation and flight.
How we work.
Six commitments that hold for every capture in the dataset.
Real, not rendered.
Every frame comes from a sensor in a real place. Simulation is a tool. It is not the ground truth.
Time, not snapshots.
4D means the world as it changes: motion, interaction, and consequence, not a single frozen scan.
Variety by design.
Captures are planned across places, conditions, and tasks, so a model learns the world and not one room.
Aligned across sensors.
Depth, LiDAR, RGB, and robot state share one clock and one coordinate frame.
Licensed for training.
Clear terms for commercial and research use, set with you during the preview.
Versioned releases.
Each release is numbered and documented, so a result you publish can be reproduced.