Automated product photography for fashion

An automation has two parts. The trigger says when it runs: a new garment, a new tag or a new outfit. The steps say what it makes: the shoot, the camera edits, the video. Publish it once, switch it on, and every product you add after that goes through the same steps.

Brands use it two ways. For product listings, every new product gets the same set of photos, so the catalog stays consistent without anyone doing the work. For campaigns, a tag can send a group of products through a different look.

Start a new automation and choose what starts it. The episode uses new garments, narrowed to skirts only, so a new jacket does not go through a skirt shoot. Narrow triggers keep each automation simple: one kind of product, one set of photos. For a collection you already have, a batch runs the same steps once, without a trigger.

The first step is the shoot itself: the art direction, the model and the image model. Every skirt that comes through uses the same three, which is what makes the photos consistent from one product to the next.

After the shoot, add camera edits on the full-body photo. The episode adds four: the lower body, a detail, the side and the back. Each edit keeps the model, the light and the set of the first photo.

Then add a video that moves from the front view to the back view. It is made from the photos the automation just produced, so it matches them.

Before you publish, the canvas shows what one run costs: here, 41 credits per skirt, or $4.10. That is the price of every new skirt from now on, so you know what a collection will cost before it arrives. Credits are $0.10 each, with no subscription.

Publish the automation and switch it on.

Add two skirts. The automation picks them up, and the photos and the video arrive in Generations and on each skirt's own page. Nobody has to open a shoot or press generate.

The same automation can be built from one request in plain words: "every new skirt gets a full-body photo, the back and a video." The agent drafts it on the canvas. A second request changes the draft, here adding the art direction. You review the draft, check the cost, and publish it like any other automation.