AnimateDiff Evolved 🧩 Workflow Open source
Improved AnimateDiff nodes for animated clips from Stable Diffusion in ComfyUI
- GitHub stars
- 3.5k
- Stars this week
- –
- Forks
- 288
- Licence
- Apache-2.0
- Last push
- 2026-07-28
- Maintainer
- Kosinkadink
cd ComfyUI/custom_nodes && git clone https://github.com/Kosinkadink/ComfyUI-AnimateDiff-EvolvedThird-party workflows & skills run with your permissions and API keys. Read the source before installing, and prefer pinned versions.
Works with
Good for
About AnimateDiff Evolved
What it is
AnimateDiff Evolved by Kosinkadink (Apache-2.0) is the main ComfyUI implementation of AnimateDiff, which adds motion modules to Stable Diffusion checkpoints to produce short animations. It extends the original with sliding context windows for longer clips, motion LoRAs, prompt travel, and advanced sampling options, and it is paired with VideoHelperSuite for loading and saving video. It remains the base of many stylised loop, music-visual and vid2vid workflows shared on community sites.
Key features
- Motion modules for SD 1.5 and SDXL checkpoints
- Sliding context windows for animations longer than 16 frames
- Motion LoRA support for camera moves
- Scheduling of prompts and settings over time
- Works with ControlNet and IP-Adapter for guided animation
- Extensive README with model links
How to install
Install through ComfyUI-Manager by searching "AnimateDiff Evolved" (author Kosinkadink), or clone the repo into the custom_nodes folder. Then download at least one motion module (the README links the original and fine-tuned modules) into the models folder. The README gives no one-line command, so install_cmd is left empty.
Pricing and rights
Free, Apache-2.0. Motion modules and checkpoints have their own licences; base Stable Diffusion models use CreativeML OpenRAIL licences.
Who it is for
Creators making stylised loops, music-video visuals and animated artwork from Stable Diffusion models.
Verdict
Still maintained and still the standard for SD-based animation, but newer video models such as Wan and LTX give far more coherent motion. Choose it for style and control rather than realism.
Pros
- Mature, well-documented AnimateDiff implementation
- Long animations via context windows
- Works on modest GPUs
- Apache-2.0
Cons
- Older SD-based motion; less coherent than modern video models
- Many model files to download
- No one-line install command
Similar workflows & skills
All ComfyUI nodes & workflows →ComfyUI-Manager 🧩 WorkflowOpen source
Install, update and manage ComfyUI custom nodes and models from inside the UI
ComfyUI-WanVideoWrapper 🧩 WorkflowOpen source
Kijai's wrapper nodes for Wan video models and related research models in ComfyUI
ComfyUI ControlNet Auxiliary Preprocessors 🧩 WorkflowOpen source
Preprocessors that make depth, pose, edge and line maps for ControlNet in ComfyUI
ComfyUI-LTXVideo 🧩 WorkflowOpen source
Lightricks' official ComfyUI nodes and workflows for LTX-2 video with audio
ComfyUI-KJNodes 🧩 WorkflowOpen source
Kijai's grab-bag of utility nodes for masks, batches, video and workflow logic
rgthree-comfy 🧩 WorkflowOpen source
Quality-of-life nodes and UI tweaks for building and running large ComfyUI graphs