Leffa is a unified framework for controllable person image generation that enables precise manipulation of both appearance (virtual try-on) and pose (pose transfer). It learns flow fields within the attention mechanism of diffusion models to achieve better warping and feature alignment between reference and target poses. The project provides pre-trained models on DeepFashion and DressCode datasets, a HuggingFace demo, and has been integrated into ComfyUI for workflow use.