Praise for AegisFlow From Fellow Civitai member and all around cool Canuck PolarisFX:
AegisFlow is a project-oriented workflow designed for those needing to auto-generate assets for various concepts in a system that handles project/file naming automatically and keeps ideas organizable. It is sectionally arranged, and many sections can be muted while searching for just the right seed/conditioning/settings.
Then re-enable the sections to generate palettes, lineart, post-processed, and other outputs when you're ready.
It has a 22+ page (and growing) manual that explains its use clearly and professionally, and is a good resource for understanding basic SD features and principles. Check it out here.
Most items are switch-controlled; no more wiring and rewiring as if you're using an old analog synth rack.
For Example, (3) below enables or disabled the iterative upscaler.
The best way to fight back against the bugs is to check out the 22-page AegisFlow manual located at GitHub or get the workflow at Gumroad
Big Changes for v.a6! Now a windows installer, and only 5.5gb!!! This is a full install, minus 2 models, 2 VAEs, and 2 controlnet models.
New instructions give direct links to these extra resources, and tell you just where to place them.
Need it preconfigured on Digital Delivery?
I now have a custom installation with everything you need for sale on Gumroad for $15. This installs a custom ComfyUI environment, gives you a desktop and program files shortcut with icon, and much more.
This also allows me to create exe-based patches, which I test beforehand, so that you won't have to spend all your time fighting pythons and more time generating images.
I'm also continually extending the workflow and adding in utility workflows. The paid installer updates will include all these, while the base free workflow will not.
So far, There are four chainable workflows (and workflow templates) that Allow you to chain together a facial detailer/modifier, a creative mask and fx system, and an iterative upscaler. There's also a "saver" template that you can use to save your results, or wire it to a simple "save image" node. 
Need it preconfigured on USB?
For those that want a preconfigured, plug-n-go solution Ioffer a 256gb MicroSD card that has an out-of-box experience for those with a bit more cash than time, and also has a wide assortment of checkpoints, LoRas, etc.
It comes with a fully functioning Windows ComfyUI install, all dependencies, extra stuff, and of course the workflows, all neatly arranged and organized. I test every card before it leaves to ensure that all files copied well and that you'll be able to generate right away.
License
As of now (v.a7), this workflow is MIT licensed. You are welcome to install it, use it, extend it, or change it in any way, for any purpose that the MIT license allows. I'd prefer that you don't just download it and resell it as is, or pretend you made it, but I can't stop you either :). If I see abuse of this, I might modify the licensing but in general I want to keep the open spirit of the Ai community, while also making money for myself and family. Thanks!
Description
v.a6 updates the workflow for recent node changes, and makes small qol improvements. I've also created a simple step-by step install document to go from a base ComfyUI installation to AegisFlow-capable for those doing a DIY install.
FAQ
Comments (8)
im setting up a new bounty and a free access to the paid version of AegisFlow v.a7 for a limited time. Watching Charlie Brown Thanksgiving with the family and then will put it in place :)
Small point release update! This fizes a small folder naming issue. See here for more details
https://drive.google.com/drive/folders/1c3VuIIVXKROwo-vQ4cXyXULIEWNdJNL8?usp=drive_link
Seems like it doesn't load the Brightness & Contrast node. and yes, the brightness_contrast_ally file is in the custom_nodes folder
This was an issue in v.a4 but is not present in the current version on my end. Thanks for the report!
I'm not a fan of over-complicated workflows wich try to do too much stuff, but is still quite impressive.
.....Not a fan of windows repacks too (i use linux) but i appreciate the effort of making it the most straightforward as possible.
Your DIY install guide should also mention RealESRGAN_x2.pth and all the embeddings you used, but i still figured it out.
N.B.: ComfyUI should work on AMD cards too on windows, using the --directml flag. but it requires pytorch-directml to be installed in the venv, wich is currently available on python 3.10 but not yet 3.11.
Awesome info and I will def update the docs. The installer places the upscalers automatically, as well as lots of other supporting models). As for the repack, thanks there---you hit the nail directly, I'm attempting to make the process of getting up and running as seamless as possible for folks, and most of them are on Windows. I figure anyone on Linux is, like you, able to figure out the DIY install :)! Really appreciate it.
Documentation Updated: Finally, you can get these two controlnet models, and the 2X Upscaler model from my Google Drive directory
"And place them in the <install root>/comfyui/models/controlnet/ directory and the <install root>/comfyui/models/upscale_models directory "
RealESRGAN is mentioned in Step 7 of the DIY, but I needed it up top too. Thanks again!
In the latest installer, I just went ahead and included RealESRGAN2 since it is pretty small; one less thing for people to get and place.