whatisgithub

What is vace?

ali-vilab/vace — explained in plain English

Analysis updated 2026-06-26

3,765PythonAudience · researcherComplexity · 5/5LicenseSetup · hard

In one sentence

VACE is an Alibaba AI model that generates and edits videos, animate a still image, swap objects in a scene, extend frame edges, or transfer motion from a reference clip, all from one unified system.

Mindmap

mindmap
  root((VACE))
    Modes
      Image to Video
      Video to Video
      Masked editing
    Capabilities
      Object replacement
      Motion transfer
      Frame extension
    Tech Stack
      Python
      PyTorch
      CUDA
      Gradio
    Models
      1.3B model
      14B model
      Wan2.1 base
    Audience
      AI researchers
      Video creators
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

What do people build with it?

USE CASE 1

Animate a still image into a video clip by providing the image and a text description of the desired motion.

USE CASE 2

Replace or move an object in an existing video by drawing a mask around the region to edit.

USE CASE 3

Extend the edges of a video frame outward to create a wider shot from existing footage.

USE CASE 4

Make a character in a video follow the movements of a person in a reference clip.

What is it built with?

PythonPyTorchCUDAGradioHuggingFace

How does it compare?

ali-vilab/vacefastapi-admin/fastapi-adminmapillary/opensfm
Stars3,7653,7653,766
LanguagePythonPythonPython
Setup difficultyhardmoderatehard
Complexity5/53/54/5
Audienceresearcherdeveloperresearcher

Figures from each repo's GitHub metadata at analysis time.

How do you get it running?

Difficulty · hard Time to first run · 1day+

Requires Python 3.10, CUDA 12.4, PyTorch 2.5.1, and a recent Nvidia GPU, no CPU fallback available.

Most models are Apache-2.0 (free for commercial use), the LTX-based model uses a RAIL-M license with additional restrictions on certain uses.

So what is it?

VACE is an AI model for video creation and editing released by Alibaba's Tongyi Lab. It was accepted at ICCV 2025, one of the top computer vision research conferences. The model handles a range of video tasks from a single unified system rather than requiring separate tools for each operation. The three core modes are: generating a video from a reference image or video clip (R2V), editing an existing video (V2V), and editing a specific region of a video defined by a mask (MV2V). In practice this covers things like moving an object within a scene, replacing one object with another, animating a still image, extending the edges of a video frame outward, or making a character in a video follow the motion of a reference person. These tasks can be combined in a single pipeline. Several model versions are available. The smaller ones (1.3B parameters) run at resolutions around 480x832 pixels across roughly 80 frames. The largest model (14B parameters) targets 720x1280 resolution. All models are hosted on HuggingFace and ModelScope, and the two underlying video generation systems they build on are Wan2.1 (from Alibaba) and LTX-Video (from Lightricks). Licenses vary by model: most of the Wan-based models are Apache-2.0, while the LTX-based model carries a separate RAIL-M license. Setup requires Python 3.10, CUDA 12.4, and PyTorch 2.5.1 or later, so a reasonably recent Nvidia GPU is needed. After installing dependencies, you download the model weights and run generation from the command line by specifying a task name, a video file, and a text prompt. A Gradio-based graphical interface is also included for interactive use. Preprocessing tools for extracting depth maps, pose information, or flow data are available as an optional separate install. The codebase also includes a benchmark dataset for evaluating the model's output quality across the supported task types.

Copy-paste prompts

Prompt 1
Using VACE's R2V mode, show me the command to animate a photo of a person walking with a text prompt describing the motion.
Prompt 2
How do I use VACE's masked video editing mode to replace a specific object in a video using a drawn mask?
Prompt 3
Set up the VACE Gradio interface so I can test video generation interactively without writing command-line scripts.
Prompt 4
What are the exact hardware requirements for running the VACE 14B model and how do I download the weights from HuggingFace?
Prompt 5
Walk me through combining VACE's R2V and MV2V modes in a single pipeline to animate and then edit a video.

Frequently asked questions

What is vace?

VACE is an Alibaba AI model that generates and edits videos, animate a still image, swap objects in a scene, extend frame edges, or transfer motion from a reference clip, all from one unified system.

What language is vace written in?

Mainly Python. The stack also includes Python, PyTorch, CUDA.

What license does vace use?

Most models are Apache-2.0 (free for commercial use), the LTX-based model uses a RAIL-M license with additional restrictions on certain uses.

How hard is vace to set up?

Setup difficulty is rated hard, with roughly 1day+ to a first successful run.

Who is vace for?

Mainly researcher.

Open on GitHub → Ask about another repo

This repo across BitVibe Labs

Verify against the repo before relying on details.