Delivery format

Rerun .rrd delivery format for egocentric episodes

One recording per episode with every stream on a single timeline to inspect visually.

By the Firsthand capture teamLast updated September 23, 2026

Short answer

A Rerun .rrd recording puts every stream on one timeline — camera, hands, point cloud and a sync-error plot. It is not a training format; it is how we recommend a new team first meets the dataset, before writing a loader, so you can see the alignment with your own eyes.

Field mapping

How Firsthand fields map into this format

Container: Single .rrd recording per episode
Entities: world/head_cam, world/hands, world/points, plots/sync_error
Timeline: All streams on one shared nanosecond timeline
Open with: rerun EP_0117_KITCHEN_A.rrd
Use case: Look at the data before you write a loader

Loading it

shell

pip install rerun-sdk

# open the recording in the Rerun viewer
rerun EP_0117_KITCHEN_A.rrd

# every stream shares one timeline: scrub to any moment and the
# camera, hand skeleton, point cloud and sync-error plot stay aligned

Gotchas

What to watch for

This is for inspection, not training — pull RLDS, LeRobot or WebDataset for a training loop.
Large recordings stream best from local disk; download before opening.
The plots/sync_error entity shows the measured residual per frame against the 2.0 ms ceiling.

Get a Rerun .rrd sample.

40 episodes across 4 environments, delivered in the format your stack already reads.

Get sample data →All formats