🎬 Desaign WAN2.2 Unified – Version Beta
This is the Beta release of our all-in-one video generation workflow powered by the WAN2.2-Rapid-AllInOne model — now updated with enhanced capabilities, better performance, and simplified usability.
🔄 What’s New in the Beta Version:
✅ Text-to-Image support added: You can now start from a prompt directly, no base image required.
⚡ Simplified setup with Fast Group Bypasser: Activate or skip components instantly for a smoother user experience.
🔗 Updated to WAN2.2 v4: The workflow now links to the latest version (v4) of the WAN2.2 model by Phr00t, with improved adherence and visual quality.
We strongly recommend switching to these updated models — they perform better across the board.
📊 Performance Notes:
All preview videos were generated at 640×480, around 40 seconds long, on an NVIDIA 3090 Ti (24 GB VRAM).
First and last frame (FLF) is currently not available due to technical constraints in this release — an update will be pushed once this feature is possible with the current setup.
🎥 Desaign WAN2.2 Unified ALPHA – Streamlined Power for Video GenAI
Introducing our latest workflow: Desaign WAN2.2 Unified — a one-stop solution built to harness the full capabilities of the WAN2.2-Rapid-AllInOne model by Phr00t.
This tool is designed to make image-to-video generation effortless, fast, and visually striking — perfect for creators who want speed without sacrificing quality.
🔧 Key Features:
All-In-One WAN2.2 Integration: Includes WAN 2.1 VAE, umt5xxl CLIP, and Lightx2v + PUSA LoRAs — already embedded in the workflow for plug-and-play usage.
Automatic Image Resizing: The input image is automatically resized vertically to ensure ideal video framing.
Simple Duration Control: A single “Duration (s)” slider replaces frame count math. Tested up to 7 seconds with stable results.
Built-in Video Upscaling: Videos are upscaled 1.5x (Resize) and then 2x with RealESRGAN, for crisp, high-res results.
Optional RIFE Interpolation: Generate smooth in-between frames for fluid motion with a toggle.
Final Frame Export: Automatically saves the last frame of your animation as a separate image — ideal for previews or thumbnails.
📦 Zero Configuration Required
Everything’s ready out of the box. No need to download extra models — it's all linked and embedded in the workflow.
🧪 Recommended Usage
Perfect for stylized and semi-realistic animations. The workflow is stable up to 7 seconds; longer durations may vary based on your system.
🧠 Part of the Desaign Tool Suite
Focused on clarity, efficiency, and creative freedom.
💬 Need help or want to share your results?
Join our community on Discord:
👉 https://discord.gg/82W94bqxzP
We’re happy to assist, showcase your work, and collaborate on what's next.
Description
FAQ
Comments (6)
Love your workflow so far! Thank you much.
Cant you make a text to video one as well?
thanks a lot, i'm glad you had fun, i will update it to include txt2vid aswell
Thank you! It perfectly worked on 12 GB VRAM / 16GB System RAM. Thank you for your efforts and time.
i think it even runs for 8gb Vram, i'm glad you liked the workflow
The approach seems legit. I however discovered for me 2 flaws. It seems like there is always a delay in the animations. Like it needs a sec or 2 to get going. It also does not seem to follow the prompts very well. Also the resolution or sharp output is not there. This could be because of the upscaler and might not be a thing without it. I really want this to work because it is so much easier to work with just one model.
Best working workflow I've used, can't wait to out the t2v version.