Extract the complete outfit and separate blazer, top, trousers, and shoes.
Restoring your images and input…
Upload one model outfit photo to scan garments and hats, select the full outfit and individual items, and generate ghost mannequin images in a batch. Preview and download from the horizontal results. Optionally create flat lays, transparent PNGs and resized assets. Review garment details and AI-filled hidden areas.
Extract the complete outfit and separate blazer, top, trousers, and shoes.
Extract the pink knit top, floral skirt, and Mary Jane shoes as separate product images.
Extract multiple items, including the cap, jacket, T-shirt, jeans, and sneakers.
A regular cutout keeps visible pixels and cannot reconstruct a complete garment after removing the person. Garment extraction generates a ghost mannequin or flat lay image and fills areas hidden by the body, such as the inner neckline, underarms and hem. These additions are generated rather than measured; check them before using the image in a listing.
No. A front photo does not provide the back construction or real measurements, so garment length, chest size, fabric composition and care labels cannot be inferred. This tool creates front product assets; sellers must check hidden details against the actual garment.
Scanning does not use generation credits. One detected garment produces one product image without a duplicate outfit image; two garments produce three images by default: the outfit and two individual items. Each selected image counts as one generation, with free allowances and credits checked for the selected count. Flat lays and regenerations are charged separately. Transparent images and padded size exports do not use additional generation credits.
Image dimensions describe the canvas aspect ratio and pixel width and height. Clothing size describes physical measurements such as shoulder width, chest and length, which the seller must measure. Aspect-ratio exports keep the garment complete without cropping its edges or inventing extra detail.
Created tasks can be restored in the same account and browser. Check the original task while it is processing. Results depend on the displayed expiry date and file retention status, so download them promptly. A network interruption does not mean the model did not run; avoid repeatedly creating new generation tasks.