AI JobsStart a program

AuraOne / Docs / Robotics Studio Open / Start here

Quickstart

Open a local dataset. Mark the failures. Export a reviewed subset.

All Robotics Studio Open docs

At a glance

Robotics Studio Open guide

Guide overview
Entry
Quickstart
Reading
5 min
Sections
2
Applies to
Applies to the current Robotics Studio Open source release and its supported JSON or JSONL manifest workflow. Check the product page and repository for the latest setup details.

01

Open the local sample

Robotics Studio Open reviews robotics episodes on your own machine. Cluster the failures. Export what passed.

Clone the repository, follow the README, and start with the bundled synthetic SO-101-style sample before loading your own supported manifest.

git clone https://github.com/auraoneai/robotics-studio-open.git
cd robotics-studio-open
# Follow the current README to start the local app

02

First reviewed subset

Start with the bundled synthetic SO-101-style sample. Scrub an episode and tag a failure. Run the sensor QA panel, then export a manifest.

The useful artifact is not a screenshot. It is the subset manifest plus the failure cluster and QA report that explain why a clip was kept or rejected.

  • Open the sample dataset.
  • Scrub the RGB, depth, and trajectory streams together. Actions too.
  • Mark a recovery, a bad contact, or a low-quality segment.
  • Export a reviewed subset manifest with checksums.