Black Forest Labs (BFL) announced the release of FLUX VTO on May 28, 2026, a virtual try-on system designed to address previous limitations in apparel e-commerce. Existing virtual try-on technology often struggled with identity preservation, garment detail degradation—such as vanishing logos or buttons—and high latency, with generation times typically ranging between 10 and 30 seconds per output. FLUX VTO is engineered to improve performance across three primary metrics for retailers. The system achieves generation speeds of under four seconds, enabling interactive shopping experiences, and is optimized to be more cost-effective for deployment across large product catalogs.

Regarding fidelity, the model preserves user identity and maintains garment accuracy, including specific textures like stitching and prints. It supports styling flexibility by allowing users to layer up to four garments simultaneously, such as shirts under jackets, while correctly managing interactions between pieces. Users can also achieve sub-second response times by self-hosting the model. To ensure safety, the system includes default content moderation, prohibiting swimwear, lingerie, and non-consensual imagery. FLUX VTO is currently available for public use via documentation and an interactive BFL Shop Demo, with integration options through FLUX MCP.

Black Forest Labs (BFL) announced the release of FLUX VTO on May 28, 2026, a virtual try-on system designed to address previous limitations in apparel e-commerce. Existing virtual try-on technology often struggled with identity preservation, garment detail degradation—such as vanishing logos or buttons—and high latency, with generation times typically ranging between 10 and 30 seconds per output. FLUX VTO is engineered to improve performance across three primary metrics for retailers. The system achieves generation speeds of under four seconds, enabling interactive shopping experiences, and is optimized to be more cost-effective for deployment across large product catalogs.

Regarding fidelity, the model preserves user identity and maintains garment accuracy, including specific textures like stitching and prints. It supports styling flexibility by allowing users to layer up to four garments simultaneously, such as shirts under jackets, while correctly managing interactions between pieces. Users can also achieve sub-second response times by self-hosting the model. To ensure safety, the system includes default content moderation, prohibiting swimwear, lingerie, and non-consensual imagery. FLUX VTO is currently available for public use via documentation and an interactive BFL Shop Demo, with integration options through FLUX MCP.