So its basically what you see is what you get, well, with a polished look?
I must have got it wrong because I thought the prompt-based input could absolutely sideline CC5 for some workflows.
Im currently downloading. I will give it an in-depth test
So its basically what you see is what you get, well, with a polished look?
I must have got it wrong because I thought the prompt-based input could absolutely sideline CC5 for some workflows.
Im currently downloading. I will give it an in-depth test
Not that much impressed for now, another thing to play and waste time.
And installing the preset models takes ages ;-(
I just aborted the install of some video presets.
Just in case, what is the best way to get rid of this?
Remove it from iClone. Found no “uninstall”…
I assume when I delete that RL_ComfyUI folder iClone will simply start without, correct?
Or would it be the best to reinstall iClone so that “patch” is removed?
The only reason I am interested in this is the NPR/Toon/anime “restyling” options
I don’t care about
"photorealism "as I use CG software to escape reality not to
re-create it in my computer.
I would like to congratulate the RL developers for starting down this road, which for some of us is of interest, especially if you want to work mostly within iClone and CC. I think positive feedback from users will only make it better. After all this is a Public Beta, not the final product. And those that have testing experience, here is your chance to contribute. I’m sure it will be much appreciated.
Is there a way to save prompt settings so you can transfer them? Would be nice to have a transfer button from the image tab to video tab
I didn’t try much… But I think my first impression is this:
I still feel much more comfortable working with Reallusion _ Unreal
In Unreal I get the render I like and TOTAL control of the scene
I always see AI attempts as doing more work, having less control, and results with lots of errors.
I think it’s to keep learning while evolving.
But when I think about creating a short or a scene… The Reallusion flow with Unreal is still second to none.
Hi toystorylab
If you need to uninstall the plugin and RL_ComfyUI, you can do so by going to Apps & Features in your Windows system and uninstalling the AI Render plugin. Once both iClone and Character Creator AI Render are uninstalled, RL_ComfyUI will also be removed.
If you need to revert to the original versions 8.54 and 4.54, simply reinstall both main programs through Reallusion Hub.
However, some folders in RL_ComfyUI may remain, as they could contain user-customized content and will need to be deleted manually.
Installation was smooth and followed the tutorial without issues. One thing to note—the model file sizes are quite large, so expect some loading time.
So far, I’ve only tested the image generation feature. It runs smoothly within iClone 8 and integrates nicely as part of the workflow. The output quality really depends on a few key factors: your original image, lighting, character design, and shot composition. That said, the results are quite impressive for a first try.
I can definitely see the potential as the AI develops further—especially when paired with a high-end setup. If video generation and consistent character output become more stable, this could become a great alternative to traditional toon-shading. It might even reduce the need to finalize in UE5, especially if this look can be achieved consistently in future updates.
Set-up:
Intel i7-8700@3.20GHz
32GB RAM
RTX 3060 12GB VRAM
Thanx for clarification!
Same here. I achieved some very interesting character renders that were certainly different, but not better. I believe that finer control centers on proper depth mapping and creating additional custom prompting, but first pass, still low on the learning curve, I was a little letdown, especially when I noted the infamous “Token” system. I will say that I’ve been using a similar workflow using Reallusion for an image feed to OpenArt for some time. It does have great possibilities, but for now, I’ll take the “free” renders and highly controlled output of the existing render system. At this stage, it’s challenging for me to make the visual handshake between the “preview” in iClone and the AI Render Output, however the time savings in compositing, tracking background images, achieving realism, and color correction cannot be denied.
Hello there,
Is it possible to render only the two technical map: Normal map & Depth map,
without rendering the complete final AI Render Output?
IMO, the skin and hair are better but the outfit is worse.