Join our Discord community for any help:
https://discord.gg/RnUHPzTkXH
comfyui-scail2-infinity
A single all-in-one ComfyUI node — SCAIL-2 Infinity (auto window) — that turns the core WanSCAILToVideo chunked workflow into one node that generates arbitrarily long SCAIL-2 video automatically.
Project Link: https://github.com/collbroGTR/comfyui-scail2-infinity
How to install?
Download and extract the .zip, you will find 2 folders, one has ComfyUI custom node, and the other one contains the workflow. Copy the custom node folder to your ComfyUI/custom_nodes directory. After that, open the workflow in ComfyUI and start using the node.
Description
FAQ
Comments (23)
A simple, high-quality workflow. Congratulations. The final output turned out fine. Is 704x1280 the only resolution available? I couldn't find an option to change the resolution. Am I to understand that this is the only resolution?
You can actually change the resolution, 704x1280(0.90MP) is the most stable one. To change the resolution check on the tracking section, there is a "Resize Image/Mask" node with a Megapixel option. The default is 0.90 which is 704x1280 on 9:16. Try changing it to 0.40, that way you should get something like 480x832
@coolstrad How would I go about changing the aspect ratio to 16:9. When I load a video it gets cutoff & remains in 9:16
@vhanjpybbzjhriosyv661 Thats probably happening because your image is 9:16. Try turning your reference image into 16:9 and it should work. If its something else then let me know.
Also, you can force change the load video resolution by manually adjusting it on the VHS video loader node.
@coolstrad TY. Looks like editing the reference image to a 16:9 format did the trick.
"Also, you can force change the load video resolution by manually adjusting it on the VHS video loader node."
I'm still new to this. Is it the first most left VHS node because when I unhook the height & width inputs it locks me out of adjusting those values
link to discord is not working for me, can you dm?
How do you get this to work with Lora? I tried putting them in the power loader but it didn't do anything. Also, does it only work with 2.1 lora or wan 2.2 as well?
please make an update with multi-references. It make video with turn around poses or keep details consistent.
Can't Scale 2 make porn?
img2video Model.
whatever image as input it will do the magic. it won't analyze the input image (AFAIK)
if u want more than basic, maybe try with the modified Text Encoder and LoRA
@Yathukku Woow, so that’s why nobody shows how to do it here! It feels like everything gets low-key hidden when it comes to this kind of stuff.
Thanks a lot for the tips, I’m gonna go look for a better text encoder and some LoRAs now. Appreciate it!
result video sometimes a glitch appears for a few seconds, Very annoying
o copied the custom node inside my folder but comfyui doesnt install it, why?
[INFO] 0.0 seconds (IMPORT FAILED): C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-scail2-infinity
i fixed updating comfyui but now there is a problem with the node comfyui_nvidia_rtx_nodes that seems incompatible with python 3.13, maybe...
@samantha i guess you should try installing comfyui environment with 3.10.17 for example
@wAFON yes i fixed now
New problem with the mask, in the mask there is a background black and the person is blue, but i noticed the final output rebuild completly the background and doesnt leave the original one also if seems identical in reality is rebuilted from scratch, in fact i tried to add an image with a small difference in the background and the final output had the modification i did also if is not covered by the blue mask... so what's the purpose of the blue mask if the final output is always all rebuilted?
It works pretty good but what should I put in Scale in RTX Video Super Resolution node? 2 for example doesnt work
If you set the mask the background seems the same but in reality it continues to change in fact scail2 redraw the entire video not only the mask, how can i let the background perfectly identic to the original without that decompression effect?
I had been looking for a way to convert the full videos instead of just few secs and this works perfectly, man. Thank you so much!! :)
Edit - I used GGUFs, SCAIL-2 and Text Encoder GGUFs work fine after replacing the nodes with GGUF loaders.
can u share it with me im having difficulty running it on my pc
Getting some grey-ish frames after every 81 frames (something framebucket blabla).
But cant get a solution for that roblem :/