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
Comfyui-WhiteRabbit
+ packaging, torchlancSage Attention (optional)
Upscaler model (optional)
Features
🖼️ I2V
🧩 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
🎭 NAG - Negative prompting with CFG1)
🪄 Interpolation feature
🔖 Bookmark-Shortcuts - with number keys
Changes / Fixes
Added resolutions
QoL changes
Quality optimizations
Optimized workflow sequence process
Changed loader logic on new native approach
gguf error
Fixed some wires
FAQ
Comments (20)
Amazing workflow, thank you. Noob question. When I use first and last frame, even with interpolation, it seems as though the resulting video always has a little crossfade effect at the very end. In other words, I provide a first and last frame. The start of the video looks great, all the way until the end, but then it weirdly fades or morphs into the last frame I provided. As if it's fading into the provided last frame, rather than smoothly just ending in the last frame. Is there a way to make the entire video smooth, rather than having it crossfade into the last frame? Thanks for the help.
This might be a bug, maybe I made a mistake ^^ I'm working an a next release.
@darksidewalker Thanks my friend. Even so, know you do fantastic work, and it's highly appreciated.
Already new version, nice. :) The FLF section group wasn't upscaling on the 1.6 workflow. Will try and see if 1.7 works.
Was not?
@darksidewalker I overlooked it when I put in a watermark node, its working :)
I am having issue with Interpolator cutting the last few frames and ruins a perfect loop, not a workflow issue. I inspect the frame info on right click to make sure, and there are a few missing. The multiplier is not accurate. Maybe there is a node pad some extra frames at the end before interpolating, or I have to do it manually in editor. Already released 2.3, you work fast.
I mean they just keep getting better and better bro!
Just having a look at V2.0 of your workflow.
Is Sigma Shift not required for the FLF2V section?
I've just noticed Sigma Shift is only connected to the I2V section
I maybe was to tired to recognise, this could be a bug :) I'll have a look!
@darksidewalker Yeah I wasn't sure if it was an error, or intended for FLF2V as I haven't really tried much looping stuff before.
It does seem like the weird color flashes are back at the end of the loop. However this seems to have slower movement, I guess due to the high/low sampler split.
I didn't notice that on my test, but I'll have a look!
I did 10 test samples (low res) now and didnt see a color flash ... but I connected the sigma nodes to the flf, I did forget this by mistake on the upload. I'll up a fixed version asap.
I got a flash using a checkpoint that was not merged by darksidewalker. If you use midnightflirt it works as intended. If you're not looping, other checkpoints work fine too.
@clumgufnu 😏yeah! As intended! 🌩️ Don't you dare not to use my sexy models! ... Jokes aside, I have no idea what would cause that. My workflow is not only usable with my checkpoint. I'm on v2.2, do clean some things up and a nice addition!
for the moe ksampler - have you tried this fork?
https://github.com/GalaxyTimeMachine/ComfyUI-WanMoeKSampler
it even has cfg scaling %, i saw you removed it from the updated workflow as of today, i had better success with this fork than the og moe
Thanks for the tip, I'll test that!
As far as my samples goes for now:
-> Same cfg, stepped cfg and same resolution.
I can not spot any differences to 2*Ksampler with 50% steps.
The prompt adherence is the same regardless if I choose cfg 1 or 1.5.
It works slightly better as the original MoE sampler, but I did too few samples to really spot a measurable diff.
I'm not sure if there is any advantage over just 2*K, except it will automatically switch on noise-ratio.
Did you got any advantage from this MoE-K?
@darksidewalker Im not sure if I’d say advantage honestly. It just seemed to be better than the original moe sampler - compared to your 2.0 workflow I didn’t seem to notice a real difference except I use sage attn 2.2 with a 5080 so I put kj nodes in between checkpoint and prompt for that with torch compile on - and the moe fork seems to render faster than traditional ksampler high/low separated.
But it could only be because of sage attn 2.2. I have sageattn3 but nothing supports it yet so it just cuda errors out.
Do you use sage attn 2.2? The torch compile function on the kj patcher supports it on 2.2+ and the moe fork seems to like it.
My comment might of got yeeted it seems.
I only notice that since I use sage attn 2.2 with my 5080 with torch compile using kj sage patcher node between checkpoint and prompt - it just seems to like it and renders slightly faster.
In terms of prompt adherence etc - I’m not sure because your newest checkpoints allowed me to test a bunch of new things so I’m dicking around a lot.