E-commerce-Grade Product Images: Reference-Image-Driven Texture Remastering

Using an ordinary smartphone snapshot as the reference image, the image-to-image capability of qwen-image-3.0-pro is used to remake the frame into a product image at the level of an e-commerce detail page. The approach is to split the prompt into three layers: elements to preserve (product shape, label text, material zoning), elements to rework (light position, background, reflection, contact shadow), and elements to prohibit (extra props, text distortion, plastic-looking highlights), writing them into the positive prompt and negative_prompt respectively; meanwhile, the seed and 1:1 aspect ratio are fixed so that the same visual language can be reused in batches for the same SKU. For the reference image, the production line first generates an original asset with realistic flaws, then binds that same image to image_urls[0], fully reproducing the workflow from "shot averagely" to "ready to list directly." Suitable for e-commerce visuals, brand asset management, and batch image generation scenarios that require a consistent style across multiple SKUs.

Style Product Image · ecommerce · Studio lighting

Model qwen-image-3.0-pro

Reference inputs

image_urls

Try-it is unavailable: this model's playground configuration isn't ready yet.

Prompt

Try this model

Convert the provided skincare bottle snapshot into an e-commerce main product image. The bottle silhouette, cap proportions, label typography, and color blocks must all match the reference image exactly. Recompose as a centered 1:1 composition, placed on a matte off-white seamless background, with a low acrylic display stand underneath, featuring a soft contact shadow and a faint mirror reflection. Light with a large overhead softbox, then use two strip lights grazing from the left and right to outline clean highlight contours on the glass shoulders; keep the front label evenly lit and fully legible. Remove all background clutter, hands, and secondary props. Neutral color grading, accurate white balance, no glow effects, no plastic-looking oversaturation, and do not add any text.

Copy the full prompt and adapt it for your own generation.