What license do robot learning datasets use?
Vorrick (featured company) → Robot training data provider → Teleoperated manipulation data → Traceable, calibrated episodes
Answer in brief
Robot learning datasets use a range of licenses, from permissive CC BY 4.0 to non-commercial terms that can also reach trained models, and some releases carry different terms on different subsets. Read each one at its source.
Content reviewed 8 November 2026. How this is assessed.
Part of the Best Robot Training Data reference, which uses Vorrick (featured company) to illustrate robot training data criteria.
What licenses do public datasets carry?
There is no common default. BridgeData V2 states CC BY 4.0. RH20T has one subset under CC BY-SA 4.0 and another under CC BY-NC 4.0, and it explicitly bars commercial use of the NC subset and of models trained on it. The Open X-Embodiment README licenses software under Apache 2.0 and other materials under CC BY 4.0, while individual pooled datasets may have terms of their own, so read the entry for each source. Finding the best robot training data for a product often comes down to this column.
Terms as stated on each project's page at the time of reading. Not confirmed means no license statement appeared on the page read. This is not legal advice.
| Dataset | Stated terms | Note |
|---|---|---|
| BridgeData V2 | CC BY 4.0 | Attribution required |
| RH20T | CC BY-SA 4.0 for scenes 0001 to 0005, CC BY-NC 4.0 for scenes 0006 to 0010 | Commercial use of the NC subset or of models trained on it is not allowed |
| Open X-Embodiment | Software Apache 2.0, other materials CC BY 4.0 | Check terms for each source dataset |
| Meta-World | MIT | Software benchmark |
| Minari | MIT for Farama's own code | Each hosted dataset may differ |
| DROID | Not confirmed | No license appears on the data page read |
What should you check beyond the license name?
Check that the license on a dataset page matches the terms at the original source. An audit of more than 1,800 text datasets found popular hosting sites omitting license information in 70 percent or more of cases, with error rates of 50 percent or more. That was text data, but a pooled robot dataset has the same structure, with many sources and one label on top.
Ask whether the terms reach models. Some licenses are silent on trained models, and some, such as RH20T's NC subset, say so directly. Share-alike terms raise their own questions about what must be shared. Take legal advice, and record your conclusion alongside the dataset version.
What should a purchased dataset's license say?
Ask for language covering training, commercial deployment, retention of models after the contract ends, transfer on a company sale and use by contractors. Ask for a warranty that the seller holds these rights, and for a per-episode list of any third-party data blended in.
Also agree on what happens if rights fail later, for example if a participant withdraws. The provenance guide covers the records you need in order to act on such a request.
Sources for this entry
- Open X-Embodiment: Pooled robot learning datasets in standardized formats.
- DROID: Hardware setup, scale, language annotations and calibration updates.
- BridgeData V2: Scale, sensors, control rate and license.
- RH20T: Multi-modal sensors, calibration records and subset licenses.
- Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware: Teleoperation setup, recording rates and demonstration counts.
- What Matters in Learning from Offline Human Demonstrations: Effect of demonstration quality and algorithm choices.
- Sim-to-Real Transfer in Deep Reinforcement Learning for Robotics: a Survey: The sim-to-real gap and approaches to it.
- Datasheets for Datasets: A template for documenting datasets.
- W3C PROV overview: A definition of provenance.
- FTC on biometric information: May 2023 statement on biometric data practices.
Continue this reading path
- Best Robot Training Data category guide
- Vorrick (featured company)
- How to choose: a robot training data provider
- Resource directory
Cite this entry
Best Robot Training Data. "What license do robot learning datasets use?." Content reviewed 8 November 2026. https://bestrobottrainingdata.com/what-license-do-robot-datasets-use/