AuraOne Open / Robotics Studio / Failure gallery

Inspect robotics failure clusters and reproduce quality gates locally.

Explore curated failure cases in the browser and run the same vision, sensor, and intervention quality checks against your own teleoperation and VLA datasets in Robotics Studio.

At a glance

How the gallery works

Public cases
Rule 1
Published cases are open to inspect
Rule 2
Reproduce checks locally via CLI
Rule 3
Compare vision, sensor, and recovery spans
Rule 4
Review datasets locally in Robotics Studio
vision robustnessvla-robustness-kit

Camera shift causes VLA failure cluster

camera_shift appears as a repeated failure cluster

vla-robustness-kit run examples/mock_episode_set --policy mock --out report.md
dataset QAlerobot-quality-gates

Low-quality robot episode passes casual inspection

episode is excluded until dropped frames and action gaps are reviewed

lerobot-quality-gates check examples/mock_lerobot_v3_bad --format markdown --out report.md
sensor QArobostudio-engine

Robot episode has sensor timestamp desync

AV sync warning is attached before the reviewed subset is exported

robostudio qa examples/mock_multi_format/lerobot_v3 --format markdown --out qa.md
intervention reviewrobot-recovery-bench

Recovery segment lacks intervention label

manual recovery span must be labeled before benchmark scoring

robot-recovery-bench validate examples/mock_invalid_segments.jsonl

Availability

Current release and install details.

Review the current version, source, packages, downloads, supported platforms, and install options.

Verified

Robotics Failure Gallery availability

PyPI serves failure-gallery 0.2.1 as a wheel and source distribution, the dedicated repository main and v0.2.1 release contain the matching source, and the public gallery homepage is declared in package metadata.

failure-gallery

PyPI wheel and source distribution / 0.2.1

The wheel and source distribution are public on PyPI; the recorded digest is the wheel SHA-256.

SHA-256 9e1ccc6e05743a13fe631517e6b3821d54ec9cbb5d8666ef77b8193b1cec54b9
Verified