My ComfyUI workflows for using Wan 2.2
This workflows are used by me to create my art.
They are optimized for my checkpoints and created of my latest knowledge to enhance the outcome.
"If this workflow leveled up your day, I'd purr-eciate a like! 😻"
Versions & Information👇👇👇👇👇👇👇👇
👉 Please read below and the file descriptions "About this version" for more info's.
🎬🐈Click me! HowTo + Video
💡Some WAN 2.2 versions use high+low checkpoints, other like S2V use only a single checkpoint - make sure to read the descriptions and use the correct checkpoints.
⚠️ Do not use the workflows with the "Nodes 2.0 beta" from ComfyUi or it will mess up things.
👇👇👇👇👇👇👇👇
What you get from the comfy workflows:
♨️ Easy controls
✅ As less as possible dependencies
🪧 Detailed documentation
⛓️ Highly automatic logic
✨ Optimized results
🎬 Fully automated resolution logic
🔖 Bookmark-Shortcuts with number keys
Types of workflows
FastFidelity C-AiO
🖼️ I2V and FLF2V
🧩 Automatic aspect‑ratio calculation and fitting
✨ Multiple Upscalers
Torchlanc (very fast, color correct, low VRAM)
Upscale with Model (additional detail, high quality)
RTX Super Resolution (ultra fast, very accurate)
🤝 Video resolution matching - Fully automatic scaling and resolution calculations
🕠 Length automation - Fully automatic calculation of frame count
🫥 Watermark option
🧮 Color match feature
👾 MiniMeme feature - Create small gif's
🎭 NAG - Negative prompting with CFG1
🪄 Interpolation feature
🌀 Perfect loop feature
🃏 Last Frame Extraction
🔖 Bookmark-Shortcuts - with number keys
FastFidelity C-SVI
🖼️ SVI (SVI 2.0 up to 10 samplers)
🧩 Automatic aspect‑ratio calculation and fitting
✨ Ultra‑fast, color‑correct upscaling (torchlanc) OR Upscale with Model
🤝 Video resolution matching - Fully automatic scaling and resolution calculations
🕠 Length automation - Fully automatic calculation of frame count
🫥 Watermark support
🪄 Interpolation feature
🔖 Bookmark-Shortcuts - with number keys
Swarm Basic
Absolute basic SwarmUI preset
Not recommended over ComfyUI for video generation
Backend Test
If this does not create the example video your ComfyUI backend is broken.
If this works, but your other workflow not, the other workflow is broken or missing dependencies
🩻 Known issues and advice's
⚠️ Some workflows may set on webp av1 encoding (VHS node) - If your computer/setup missing drivers use any other like H265 or H264!
Install ffmpeg!
Update Comfyui and custom_nodes!
Update pytorch 2.9+cu128 or higher
Make sure to read where files/models should be placed inside the workflow
Check if the filepath for model/clip/vae match your system like Linux/Windows
The plugin ComfyUI-DD-Translation can break node connection (avoid)
All older Versions are available inside my GitHub Repo.
Spacial thanks to @Abyss_Games for a really good idea and help with the "loop"!
I got some really good ideas from @Gladas workflows!
YOU are responsible for outputs as always! If you make ToS violating content and I get aware I WILL report this.
Description
## Requirements
- https://github.com/Kosinkadink/ComfyUI-VideoHelperSuite
- https://github.com/rgthree/rgthree-comfy
- https://github.com/Artificial-Sweetener/comfyui-WhiteRabbit
packaging, torchlanc
- https://github.com/kijai/ComfyUI-KJNodes
- https://github.com/yolain/ComfyUI-Easy-Use
- https://github.com/princepainter/ComfyUI-PainterI2V
---
## Features
- 🧩 Automatic aspect‑ratio fitting
- 🖼️ Monitor‑resolution selective upscaling
- ✨ Ultra‑fast, color‑correct upscaling (torchlanc)
- 🔖 Bookmark-Shortcuts with number keys
- Video resolution always matches optimal Div16 WAN 2.2 resolutions
- 💫**Faster motions** through PainterI2V
---
## Changes
- Changed upscale logic
- Changed Interpolation logic
- Changed resolution matching
- Optimized layout
- Gathered all settings
- Changed WAnImage2Video to PainterI2V
## Fixes
- none
FAQ
Comments (24)
Amazing work brother ! <3
This WF helped me to run the model very smoothly :)
I like the latest version. However...
I could not find the fast groups bypasser so I had to add it myself.
Allot of people use 4 step lightening Loras and I believe even your latest checkpoints have them built in. Might I suggest incorporating the new PainterI2V node instead of the traditional WanImage2Video node. It allows you to have more control over the speed and motion. I incorporated it into your workflow and it works great. Plus, it does not conflict with any other nodes.
https://www.youtube.com/watch?v=lNRncdll5Rs
It is inside the EX version, but I recognized some odd behavior with Loras, the node is not universal compatible.
The bypass of groups is manages in the settings group, I made it more automated and as a switch there
It's strange that when the generation is done, it doesn't show the video preview unless you switch to h264. Is this intentional? Last version also had this problem btw, but at least you get the last frame, but now you get nothing so a newbie might not know it generated anything at all.
This must be a browser thing, on my end everything is displayed, except you disable it inside the node.
Been trying out this workflow. When I tried to upload a video to Civitai, I got an error "Failed to load video. This may indicate that the file is poorly encoded for use on the web." Any ideas?
I compared it to another workflow and I think it's just the default video codec options. Seems better using H264.
The WF use standard H265 encoding. Because it is superior to H264. H265 works fine with mp4 on civitai, if the video is not encoded this might be a problem of your browser or your codecs on your computer.
it's fantastic, thank you!!!
how to upscaling in the latest workflow?
It is upscaled automatically to twice the resolution to maintain aspect, color and sharpness. If you want more than that, you could adjust the "math" nodes to 3 or 4 ... Maybe I could set a setting for that ... but I find 2x a good value to get consistent results.
@darksidewalker i see, thank you
Just fantastic. Not just your models but this workflow. Have it working on a 3060 ti (8GB). I do lower frame count to 49 but this puts wan 2.2 in my hands and I love it.
As an aside I have 32GB of system ram and a 48GB swapfile (Linux). I specifically enable the swapfile for image and video generation.
Increase main memory to avoid using swap files!
Just eliminating the 48GB swap file should cut generation time roughly in half.
It rapidly consumes SSD lifespan (TBW), which is really bad news.
128GB is best, but if you can't afford it, switch to a lighter workflow using models like Q3_K_M.gguf.
Love the WF, the only issue I'm seeing is that when generating vids longer than 5 seconds, they naturally try to get back to their starting position.
For example, with a 5 second vid a woman will take her shirt off. With a 7 second vid a woman will take her shirt off then attempt to put it back on. This is, obviously, with FLF and Exact Loop off, using I2V only.
Also, for whatever reason, in the latest WF I dl'ed (C-AiO-1.0), the Video Combine node no longer shows a preview video? It still saves it, but the preview never shows up in comfy?
With the undressing part you are right! That is a WAN thing, ether you fill more time describing things so the model does not have to guess or it will just guess up things turning around to its start, as you noticed!
The problem with the preview not showing ... hm ... I could not reproduce that ... I'm wondering, on my part everything shows up.
But if you prompt for dressing up, it will refuse! Darn! Such a nasty AI! ~ 😅 In fact I found it more difficult to dress things up!
@darksidewalker interesting! I only recently switched to 2.2 so didn't know it naturally did that.
An update and a reload/refresh seems to have fixed the preview issue, sorry for jumping the gun!
its standard for almost all ai vids that are over 89 frames...it boomerangs back to starter image...so either keep em around that length, or use it as part of the video. like if theres a bottle on a table and you prompt her to pick it up, drink from it and set it back down...use 150 frames so it uses the boomerang effect to put the bottle back where it was.
@leonqubick591 Yeah I know, because the base is trained on 81 frames/5s.
Looks like good, but function to disable automatic upscale is required. =\
Than disable it?
I must have something wrong in my flow. Mine is giving me the error "The size of tensor a (81) must match the size of tensor b (16) at non-singleton dimension 1"
I have absolutely no idea what this means...