Dataset profile
Open X-Embodiment
Open X-Embodiment pools real robot trajectories from many labs to support cross-robot transfer and RT-X-style model training. It is a useful benchmark for robotics buyers comparing robot-native action data with human demonstration data.
Overview
Open X-Embodiment pools real robot trajectories from many labs to support cross-robot transfer and RT-X-style model training. It is a useful benchmark for robotics buyers comparing robot-native action data with human demonstration data.
What's included
- Robot trajectories from many datasets
- Multiple robot embodiments and action spaces
- Common robot-learning data mixtures
Tasks / activities
Modalities
Collection methodology
The dataset aggregates existing robot datasets contributed by research labs and normalizes them for cross-embodiment training.
Potential applications
- Cross-embodiment pretraining
- Robot foundation models
- Action-space transfer
- Multi-task manipulation
Access & licensing
This is a third-party public/research dataset. We do not own or distribute this dataset. Access and usage are governed by the original publisher's terms.
The pooled datasets carry source-specific terms; users should review the original dataset licenses.
Limitations / opportunities for custom data
Custom healthcare data can add expert human demonstrations, clinical task taxonomies, and care-setting constraints absent from broad robot pools.
Need data beyond this benchmark?