

Before and After
Freegma is a local AI generation studio that unifies multiple ComfyUI workflows into a single, clean interface. Text to image, image to 3D, image to video, and lip sync — all running on your GPU with zero cloud dependency.
The problem was simple: I had too many ComfyUI workflows and kept switching between them. So I built Freegma to put them all in one place. It runs locally through a Cloudflare tunnel, so you can access it from your phone or share it with anyone.
The UI/UX was redesigned using Figma Make, with custom animations.
Coming soon: LoRA-trained character consistency for AI-generated ads. Describe a character and an action, feed in a photo of a clothing item (shirt, pants, etc.), and the character wears that exact piece. Clothing LoRA is already trained and tested locally. Character LoRA for ad generation is currently in training.
0 likes · 19 views · Jun 14, 2026