Generating AI product photos for a Shopify catalog is a four-step loop — pick the products, choose one look, preview the first four images, then publish or cancel — because bulk generation without a preview step is how a catalog gets ruined faster than it gets better.
The workflow below assumes a real store: hundreds of variants, a mix of photography quality, and originals you cannot afford to lose. Each step exists to keep the run reversible.
Step 1: choose the look before you choose the products
Three looks cover almost every Shopify use case, and they are not interchangeable:
| Look | Where it works | Watch out for |
|---|---|---|
| Clean white background | Product grids, listing consistency, channel requirements | Boring as a hero image; products need clear edges |
| Styled scene | Collection banners, lifestyle sections, social | A scene that fights your brand palette looks like stock |
| On-model | Apparel, accessories, anything worn or held | The people must be synthetic or licensed — never a random face |
One look per run is a feature, not a limitation: mixing looks across a collection is what makes a grid look assembled from three different stores. On-model output deserves a note of its own — model presets generated by a pipeline and labelled as synthetic never depict real people, which is the honest way to handle “who is wearing this” at catalog scale.
Step 2: select the products the run should touch
Select by collection, tag, search or your whole catalog, then freeze the selection when the run is created — what you approve at preview time is exactly what gets charged for. A frozen scope also gives you a clean unit of work: one collection per run is easy to review, easy to undo and easy to explain.
Two practical rules:
- Start with the collection that embarrasses you most. The products with the weakest photos prove the workflow’s value fastest.
- Count variants before you approve. A run over 400 variants is a real credit spend; the preview step exists so you see four real outputs before the rest is charged to.
Step 3: preview, then decide
A run generates the first four products as a preview before the rest is charged to. Look at them like a shopper would: is the product recognizable, is the background clean, does the crop keep the whole item in frame? Confirm to continue or cancel and keep the remainder of your credits.
Previews are real outputs
They cost credits like any other image, and that is the point — a preview that were free would tell you nothing about what the run produces. Four paid previews are cheaper than 400 paid regrets.
Step 4: publish next to your originals, and keep undo
Publishing appends new media beside the source images; nothing already in your catalog is replaced, resized or recolored, and the app stores Shopify file and media ids rather than your image bytes. Every run keeps a snapshot, so undo removes what the run added and restores your featured image and media order — for the whole run or one product — while the snapshot is inside its retention window.
The catalog-side hygiene still applies to generated images: keep the background family consistent within a collection, write alt text that describes the actual product, and don’t promote a generated image to featured until you have seen it at grid size.
Cost control, in numbers
The pipeline is priced per image so a run’s cost is predictable before you click: 15 free credits to start, one credit per image in every mode, HD outputs at two credits each, and a failed image refunded automatically — a provider refusal comes back as a failed, refunded item rather than a silent charge. Undo retention is 30 days on Free and Starter, 90 days on Growth and Pro.
Capyo: AI Product Photos Bulk implements exactly this loop — selection, one look, four previews, appended media, one-click undo. Run your first collection on the free plan and judge the output on your own grid. If you are still deciding what to change first, the free tools hub has the other checkers on this site.
How do I generate AI product photos for a whole Shopify catalog?
Select by collection, tag, search or all products, choose one look — scene, clean white or on-model — preview the first four images, confirm, and let the run finish in the background. Outputs are published as new media next to your originals, and the run stays undoable while its snapshot is in the retention window.
Do AI product photos replace my original images?
No. Outputs are appended as new media; the source image is never replaced, resized or recolored, and the app stores only Shopify file and media ids rather than your image bytes. Your original stays the featured image until you choose otherwise.
Should my Shopify product photos have a white background?
For grid consistency and marketplace-style requirements, yes — a clean white background keeps the collection visually uniform. Use scenes for hero and lifestyle sections instead. The rule that matters most is consistency within a collection, not one background for the entire store.
Who are the people in on-model AI photos?
They should be synthetic — model presets generated by the pipeline and labelled as such — and never real people used without a model release. In this app v1 does not accept uploaded models, so every on-model preset is generated and disclosed as synthetic.
What does bulk AI product photo generation cost?
One credit per image in every mode, HD outputs at two credits each, and failed images refunded automatically. The free plan starts with 15 credits once; paid plans add monthly credits and longer undo retention.
Free tool
Product photo cost calculator
What your product photos cost per month with your current workflow and with AI images — in euro and in hours.
Open the free toolCapyo is an independent app and is not affiliated with, endorsed by or sponsored by Shopify. "Shopify" is a trademark of Shopify Inc. Generated images are owned by you, the merchant, who is responsible for what they publish. Model presets are synthetic and do not depict real people. Capyo does not promise any sales, revenue or conversion outcome.