FLUX.2 E-commerce Product Image Prompt Paradigm: Glass Perfume Bottle Still Life

Using FLUX.2 (Black Forest Labs' text-to-image model family) as the vehicle, this demonstrates a directly reusable e-commerce-grade product image prompt structure: subject and materials → set and props → light source and lighting position → lens and depth of field → color and post-processing texture. A glass perfume bottle is chosen as the subject because it simultaneously contains three typical material challenges: transparent refraction, highlights on metal parts, and a frosted label—concentratedly testing the model's handling of reflections, shadows, and text boundaries. On the input side, only a prompt and a few non-media parameters are used (4:5 vertical composition to fit e-commerce main image slots, 2 MP resolution, png output), with no reference images, so the result is entirely determined by the text description, making it easy for developers to observe the causal relationship of "prompt change → image difference." Applicable scenarios: batch generation of e-commerce detail pages and main images, brand visual style exploration, first drafts of advertising placement assets, and preparing clean base images for subsequent image editing steps. The prompt wording deliberately maintains structured clauses, making it easy to swap the subject term and migrate to other categories (skincare, beverages, 3C accessories).