2025 SDXL-Flux / Illustrious + Programing

Current 2025 stuff.

Trying to dump windows so I'm running Pop!_OS Linux now. I can still boot to windows but I spend 99% of my time using Linux.
I miss Affinity Photo. I made some scripts for Krita but it's still ass.
I also miss Eagle.cool. Outside of that, Linux has been a a challenging but great experience.

Anyways, I'm still in love with using AI. Outside of that I have been trading on the market and helping my Mom down in Texas when I can.

I have been playing AI Roguelite and building an Infinite Academy of factions, magic, and eldritch horrors. I have never felt more creative in my life.

Shortly after getting hired at Series I leaned into AI video exploration. I made this video from a music clip from tsubi club. https://soundcloud.com/tsunamibitchclub

PNG chunk editing tool I created over the past few days with python and a lot of time in Grok / ChatGPT/ Claude.

This tool edits the image PNG chunks to implement my custom comfyui workflow. No more fuss moving from a1111/forge to comfyui.

AI checkpoint and lora gallery I created based on the A1111 / Forge gallery. I missed the magazine view to browse my models with easy access to their .json files for extra info so I just made my own with python.
I need to put this on git. :P

Current (first pass) txt2img workflow

Current (second pass) img2img workflow.

Detail pass
On larger images that need more detail in certain areas I toggle on this crop and use my html / ImageCropTool to get the x/y for where I want the extra detail. When I am happy with this detail pass I photoshop it over the prior upscale image.

After cleaning up a lot of images from an overly complicated pony to flux workflow I made a solid portrait lora I was happy with. I then mixed it with a new lora I made based on small sculpted objects.