model-wearing-productlisted

# On-Model Product Imagery On-model shots communicate fit, scale, drape and desire in a way a flat product shot can't. They're also where the most constraints stack up: likeness rights, fit accuracy, body representation rules and the model consistency problem. ## Build a model, don't license one Same argument as `ai-avatar-presenter`, and it applies more sharply here because apparel imagery gets scrutinised. | | Real model | Stock image | Generated model | |---|---|---|---| | Likeness permission | Contract and release | Licence often **excludes** synthetic use | Yours | | Consistency across catalogue | Rebooking | One image only | Unlimited | | Cost per new product | Day rate | New licence | Credits | | Appears on competitors' sites | No | **Likely** | No | **Stock portraits are frequently a licence breach as an AI reference.** Most stock licences either prohibit use as input to AI systems or prohibit depicting the model as endorsing a product. Both apply to on-model product imagery. Generate a model, save them, reuse them. ## Build the reference set ``` 1. generate_image → iterate the model at 1-9 credits until right 2. Generate 3-4 more views at the same seed, varying only the angle 3. create_brand_asset(type="character", ...) → free 4. Every product shot references the set ``` ``` generate_image( model="flux-2-pro", prompt="Full-length portrait of an adult woman, mid-thirties, " "medium build, shoulder-length dark hair, neutral " "expressi