RoboVerse

A unified ecosystem for scalable and generalizable robot learning: simulation infrastructure, task and asset packs, benchmarks, and learning pipelines.

Choose Your Documentation

MetaSim

The standalone simulation framework: installation, unified APIs, simulator backends, concepts, troubleshooting, and API reference.

Open MetaSim docs

RoboVerse

Dataset and benchmark docs, task packs, assets, RoboVerse Learn, examples, and RoboVerse-specific workflows built on MetaSim.

Open RoboVerse docs

Quick Start Tracks

Run Simulation

Use RoboVerse Data

Train Policies

Ecosystem Architecture

MetaSim

Core simulator abstraction, configuration model, state protocol, task runtime, and backend integrations.

RoboVerse Pack

Robots, scenes, objects, task packs, and benchmark assets registered as downstream MetaSim content.

RoboVerse Learn

Imitation learning, reinforcement learning, and VLA training pipelines over RoboVerse/MetaSim environments.

Citation

@misc{geng2025roboverse,
  title={RoboVerse: Towards a Unified Platform, Dataset and Benchmark for Scalable and Generalizable Robot Learning},
  author={Haoran Geng and Feishi Wang and Songlin Wei and Yuyang Li and Bangjun Wang and Boshi An and others},
  year={2025},
  eprint={2504.18904},
  archivePrefix={arXiv},
  primaryClass={cs.RO},
  url={https://arxiv.org/abs/2504.18904}
}