Product photos: one shot of your product, a full set of listings
Drop in a product photo and say how many images, which scenes and which sizes. It connects the product to every generation node as a reference and lays out hero shots, lifestyle scenes and long detail pages in one pass, without the product drifting.
This is our new insulated bottle. Make a set: one white-background hero, one each for outdoors, desk and kitchen, plus a long detail page.
What it builds on the canvas (illustration).
How it does it
It reads the product first
It reads the product image first: material, color, shape and selling points. How each shot is staged and what it emphasizes follows from that.
The product feeds every node
The product image is connected to every generation node as a reference, with its look locked in the prompt. Scenes change the background and light, not the product.
Sizes by purpose
Hero, scene, vertical and long detail images each get their own aspect ratio on the node. For another marketplace, change the ratio and regenerate.
Fix only what's wrong
If one image in the set isn't right, regenerate just that one and leave the rest alone.
Questions
How does it keep the product accurate?
The product photo is connected to every generation node as a reference, and the prompt spells out its appearance. Regenerate any single image that drifts. If exact text or logos matter, check the final images by eye.
Does it spend credits without asking?
No. Before any generation that costs credits, an approval bar lists what this batch will do, and nothing runs until you approve. Building nodes, writing prompts and arranging the canvas cost nothing.
Where do the results go?
They stay on the canvas nodes, where you can revise any single one. You can also have it write the finished files back to a folder on your computer.
Do I need to write prompts?
No. Say what you want in plain words and it breaks the request into a prompt for each node. If one isn't right, edit that node directly.