TL;DR: An experimental LoRa-based workflow by AliveAi enables text-to-image generation with outfit transfer from a reference image, available for ComfyUI.
Summary: AliveAi released Krea2, a text-to-image tool that transfers outfit styles from a reference image using a LoRa model. It provides two ComfyUI workflow options: one with higher accuracy but slower speed, and another faster but slightly less faithful. The tool is experimental, currently trained only on female outfits and may sometimes generate two people in one image.
Why it matters: This gives AI builders a practical way to control clothing in generated images via reference images, opening up fashion design and character consistency tasks. Try the workflows on ComfyUI, adjusting prompts and seeds to mitigate known issues.
Source: reddit