Composite Anything
VACE creatively provides solutions for video generation and editing within a single model, allowing users to explore diverse possibilities and streamline their workflows effectively, offering capabilities including Move-Anything, Swap-Anything, Reference-Anything, Expand-Anything, Animate-Anything and more.
A young boy rises from his chair and walks briskly to the right side of the frame towards the edge of the sun-drenched frame, as if chasing a new adventure. His eyes were bright, and the corners of his mouth were slightly upturned, revealing curiosity and excitement about the unknown...
The video shows a person riding a horse on a wide grassland. He has light purple long hair and are dressed in traditional clothing, wearing a white top and black pants. The animation modeling style gives the impression that they are engaged in some outdoor activity or performance...
The elegant lady carefully selects bags in the boutique, and she shows the charm of a mature woman in a black slim dress with a pearl necklace. Holding a vintage-inspired brown leather half-moon handbag, she is carefully observing its craftsmanship and texture. The interior of the store...
In the style of classical oil painting, the background is a river, and in the center of the picture is a mature and elegant woman, wearing a long skirt and sitting on a chair. She took the red heart-shaped sunglasses from her arms with both hands and put them on...
The video shows an old movie-style scene in retro tones, with a little penguin and a kitten having a joyous bike race. Little penguins and kittens, both dressed in orange-and-red race suits, ride vintage multi-wheeled bikes on a nostalgic dirt road flanked by spectators...
Anime-style, hot-blooded teenager in bright orange long-sleeved pants sportswear, standing on a surfboard, facing the golden sunshine in the rough sea. The teenager's short yellow hair is flying in the wind, his eyes are firm, and he has a confident smile on the corner of his mouth...
Video Rerender
VACE can perform video re-render, including content preservation, structure preservation, subject preservation, posture preservation, and motion preservation, etc. (Note: The original video is on top, and the generated video at the bottom. You can view the changes between the two by slightly hovering the mouse.)
The camera begins intimately focused on a cluster of grapevines, a close-up showcasing the ripe, plump grapes, sunlight filtering through the leaves and illuminating their amber translucence. The camera slowly moves forward and initiates a gentle upward rotation, gradually revealing the rolling hills of a vast vineyard, rows of vines stretching in neat lines towards distant hills...
In a documentary style, a row of four meerkats dances together in the African savanna at noon...
The video showcases a charming French-style café in Paris, where a lion dressed in a suit elegantly sips coffee. The lion leisurely holds a coffee cup in one hand and takes a sip to savor it. The café is elegantly decorated, with soft tones and gentle lighting illuminating the space around the lion...
A person is painting on a canvas outdoors, using a palette with various colors of paint. The person is wearing a dark blue jacket and a matching beret, and is seated on a wooden chair. The canvas depicts a landscape with a body of water and mountains in the background. The person is carefully applying...
An elegant lady is passionately playing the violin, with an entire symphony orchestra behind her...
An eagle is flying over a calm blue ocean under a clear sky. The eagle, with its brown and white feathers and yellow beak, descends towards the water, its wings spread wide. As it approaches the surface, it dives into the water, creating a splash...
Community Creativity
We're thrilled to present these amazing video cases created by our community, all made using our VACE! They truly highlight everyone's incredible creativity and the model's potential. A huge thanks to all contributors—your passion is what keeps this project vibrant and exciting. Please note that all content displayed below is independently created by community users, and their copyrights and views belong to the original authors, not representing this project's stance. If you believe any content infringes on intellectual property rights or is inappropriate, please contact us immediately. We will investigate and promptly remove it.
Reference-based Outpainting
Reference Image, Depth, and Pose-based Joint Control
Deformable Mask-based Camera Control
Reference Face and Lip-sync Consistency-based Talking Head
Reference Image-based Inpainting
Contextual Reference-based Long Video Generation
Point-based Subject Control
Contextual Reference-based Looping Video Generation
Multi-conditional Combination-based Control
Mask-based or Overlaying Reference Image-based Video Erasure
Reference Image and Condition-based Action Control
Reference Image and Mask-based Face Swapping
Geometric-based 3D Rotation Control
Trajectory-based Motion Control
Acknowledgements
We would like to express our sincere appreciation for the contributions of many colleagues for their insightful discussions, valuable suggestions, and constructive feedback, including: Yuwei Wang, Haiming Zhao, Chenwei Xie and Sheng Yao for their data contributions, and Shiwei Zhang, Tao Fang, Xiang Wang for their discussions and suggestions. Our heartfelt thanks go once again to all community creators who contributed to this project. Additionally, the template for the webpage is sourced from GoKu, thanks for its design.
BibTeX
@inproceedings{vace,
title = {VACE: All-in-One Video Creation and Editing},
author = {Jiang, Zeyinzi and Han, Zhen and Mao, Chaojie and Zhang, Jingfeng and Pan, Yulin and Liu, Yu},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision},
pages = {17191-17202},
year = {2025}
}
