Rayform vs Rerun
Open-source logging, visualisation and an experimental training dataloader for multimodal data, versus Rayform's managed processing of recordings into checked, versioned training episodes.
What Rerun does.
Rerun builds an open-source SDK and viewer for logging and visualising multimodal data over time, and a database layer for physical AI data. Recordings can be registered in a catalog where each recording is usually one episode, and an experimental dataloader presents them as PyTorch datasets. Rerun imports and exports LeRobot datasets.
They work together. Rayform can read Rerun and LeRobot data, and processed episodes can be opened in the Rerun viewer.
Where they are strong
- A fast, well-liked open-source viewer and SDK with a strong developer experience.
- Native handling of multiple timelines, 3D data and video.
- A PyTorch dataloader with random access and multi-worker loading, and LeRobot import and export.
How Rayform differs
- Rerun shows data as it was logged. Rayform estimates clock offsets between streams, checks calibration and flags episodes with dropped frames or failed tasks.
- Rayform cuts recordings into episodes at interactions and labels them; with Rerun, episodes are whatever recordings you log.
- Rayform is a managed service with versioned, hashed slices; Rerun is tooling you host and assemble yourself.
Side by side.
| Capability | Rayform | Rerun |
|---|---|---|
| Works with your own recordingsUpload from your own robots, rigs or phones. | Yes | Yes |
| Sensors and robot state alignedRGB, depth, LiDAR and joints on one clock and one coordinate frame. | Yes | Partly |
| Interactions on the timelineGrasps, placements and moved objects detected and marked. | Yes | No |
| Search across recordingsFind every grasp, scene or condition across all your data. | Yes | Partly |
| Versioned training slicesFrozen selections that always return the same data. | Yes | No |
| Export to training formatsLeRobot, RLDS or WebDataset for training; USD for simulators. | Yes | Partly |
| 3D scenes for simulationSplats and collision meshes of the place, built from your recordings. | Yes | No |
| Self-serve accessStart without a sales call. | Partly | Yes |
Choose Rerun when
- You want open-source tools to log and inspect data, and a team to build the pipeline around them.
- Interactive visualisation during development is the main job.
Choose Rayform when
- You want collection quality checks and training-ready episodes without building and running the pipeline.
- You need versioned slices that trace every checkpoint back to its data.
Sources
Based on public information as of October 19, 2026. If anything here is wrong, write to contact@rayform.world and we will fix it. Names and trademarks belong to their owners.