MotionStream: Real-Time Video Generation with Interactive Motion Controls

Joonghyuk Shin1,2 · Zhengqi Li2 · Richard Zhang2 · Jun-Yan Zhu3 · Jaesik Park1 · Eli Shechtman2 · Xun Huang2,4

1Seoul National University · 2Adobe Research · 3Carnegie Mellon University · 4Morpheus AI

Code Release

Note: The code is currently under the company's internal review for open-sourcing. While we are working on it, we have not yet received a firm timeline, and the release may or may not occur depending on the outcome of the review. Since our models, data, and codebase are based on open-source resources, we hope researchers can reproduce our results using the methodology and implementation details described in the paper. If not approved, we will use this place to share some of the evaluation data and model outputs. Feel free to open an issue or contact me ([email protected]) with any questions. We'll keep this README updated if anything changes :)

Citation

@inproceedings{shin2026motionstream,
  title     = {{MotionStream: Real-Time Video Generation with Interactive Motion Controls}},
  author    = {Shin, Joonghyuk and Li, Zhengqi and Zhang, Richard and Zhu, Jun-Yan and Park, Jaesik and Shechtman, Eli and Huang, Xun},
  booktitle = {Proceedings of the International Conference on Learning Representations (ICLR},
  year      = {2026}
}