Outfit Anyone

Outfit Anyone is an open-source virtual try-on system designed to generate realistic clothing visualizations on human subjects. Built on diffusion-based image generation, it enables users to apply any garment from a reference image onto a target person while preserving body shape, pose, identity, and fine garment details. The system supports high-resolution outputs and is engineered to maintain texture fidelity, fabric patterns, logos, and shading, making the synthesized outfits appear consistent with both the original clothing and the target image context.

Key capabilities include person-image and garment-image input, automatic garment extraction, and detail-preserving warping that adapts clothing to different poses and body types. Outfit Anyone is suitable for applications such as e-commerce virtual fitting, fashion design visualization, content creation, and research on image-based virtual try-on. It can be integrated into pipelines where batch processing of outfits and models is required, and its open-source implementation allows developers to customize or extend the model for specific domains, such as niche apparel categories or stylistic transformations. By providing controllable, high-fidelity try-on results, Outfit Anyone enables more efficient visual prototyping of outfits without the need for physical photoshoots or manual image editing.