Voost: A Unified and Scalable Diffusion Transformer for Bidirectional Virtual Try-On and Try-Off

Seungyong Lee^*   Jeong-gi Kwak<sup>*
* Equal contribution</sup>
NXN LABS

Voost jointly handles virtual try-on and try-off within a single transformer, achieving high-quality results while remaining robust to various human poses, garment categories, backgrounds, lighting conditions, and image compositions.

News

• [2025-08-19] 🚀 A public demo of Voost is now available on Hugging Face Spaces. Try it out and don’t forget to leave a 🤍 like to support us!
• [2025-08-08] 📄 Voost is now on arXiv. Visit the project page for more details and results.

Demo

🚀 A public demo (try-on & try-off) is now available on Hugging Face Spaces.
Try it out and don’t forget to leave a 🤍 like to support us!

Examples

Citation

@article{lee2025voost,
  author  =   {Seungyong Lee and Jeong-gi Kwak},
  title   =   {Voost: A Unified and Scalable Diffusion Transformer for Bidirectional Virtual Try-On and Try-Off},
  journal =   {arXiv preprint arXiv:2508.04825},
  year    =   {2025}
}

License

This project is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

For commercial use, please visit NXN Labs.

Acknowledgments

Most models and clothing images used are from internet and public datasets (VITON, DressCode). All images and brands are the property of their respective owners.