To use AI product photography for Shopify with Kive, install the app, connect a workspace and open Create with Kive from a product in Shopify admin. Generate a scene, check it against the real product, then add the approved result to that product's images. You can also import catalog photos into reusable product models for work in the full creative workspace.
Connect the store and choose your entry point
Kive's Shopify integration documentation is the reference for installation and supported actions. Before starting, confirm that the person doing the setup administers the intended workspace and can install Shopify apps. A teammate who can generate images may still lack permission to connect the store.
Install and select the workspace
- Open Kive on the Shopify App Store and select Install. Alternatively, open your avatar menu in Kive, then Settings → Integrations → Shopify → Install.
- Review Shopify's request for product and file access. The integration uses that access to read source images and add generated images back when you choose.
- Select the workspace the store belongs to and click Connect. If you administer exactly one workspace, the connection selects it automatically.
The Shopify row in Kive's settings appears only for workspace admins on paid plans. If it is missing, check the selected workspace, its plan and your role before retrying. This restriction applies to the Settings shortcut; installing from the Shopify App Store also lets you connect a free workspace you administer [1].
Choose import or embedded generation
For reusable product models, open Organize → AI models. Under Import from Shopify, find the product and click Import. The model uses up to four existing product images, and you can reference it in a prompt with its model name. Use Load more products to browse a longer catalog. A product with no images needs at least one photo added in Shopify before it can be imported [1].
For a listing update, open the product in Shopify admin and choose Create with Kive from the product actions menu. The embedded app starts with that product selected. You do not need to download the catalog image and upload it by hand for this route.
The installed app reads the catalog through Shopify's API. Import from URL is a separate option that reads a public storefront; a password-protected page or bot protection can block that route. Connecting the app avoids that public-page dependency, although account permissions and source images still matter.
Generate a scene and add the reviewed result
Inside Shopify, the embedded Kive app lets you choose a studio, aspect ratio and quality before generation. Generations use the connected workspace's credits, just as they do in Kive itself. Keep the first brief narrow enough that you can tell whether the result meets it: for example, a front-facing bottle on a neutral surface with room around the cap.
Start from the product you actually sell
Choose source photos that show the correct packaging, color and variant. A reference with an old label or a different lid gives the generator the wrong starting point. Preserve the original photos so the reviewer can compare them with the proposed output.
Pick a studio whose composition suits the listing. A tall pump bottle needs space above its dispenser; reflective jewelry needs a setting where the edges remain legible. The AI studios guide explains how reusable visual settings help repeat a direction across products.
Review before adding the image back
Compare the generated result with the original before choosing the option to add it to the Shopify product's images. Check the label, closure, proportions and visible details. A photorealistic bottle with invented text is still the wrong product image.
Each result can be downloaded or added to the selected product's images. Open advanced tools in Kive opens the full Create page when you need more control [1]. Use the editing guide if resolution needs attention, then review the edited result again.
Adding an image back changes product media. Check gallery order and any intended variant association in Shopify afterward; do not assume the generation step has configured the listing presentation. The Shopify tool comparison covers other workflows if your main need is batch background editing rather than studio scenes.
Check dimensions and the actual storefront crop
Shopify accepts product and collection images up to 5000 x 5000 pixels or 25 megapixels, with a file size smaller than 20 MB. For square images, it says 2048 x 2048 pixels usually displays best [2]. These are platform specifications and recommendations, not proof that a generated image is accurate.
Inspect the file before the gallery
| Check | What to verify |
|---|---|
| Dimensions | Read the exported file's pixel dimensions; an aspect-ratio setting alone does not establish resolution. |
| File size | Keep the upload smaller than Shopify's 20 MB limit. |
| Crop | Leave room around the product so the chosen theme crop does not hide the cap, handle or edge. |
| Small details | Inspect logos, dosage text, stitching and other identifying details at the displayed size. |
| Compression | Check that compression has not blurred lettering or created halos around a cut-out. |
Upscaling can enlarge a file, but it cannot verify that a generated detail matches the item. If lettering is wrong, return to a trustworthy source or correct the image before uploading a larger version of the mistake.
Review both the grid and the product page
Open the collection grid and product page on desktop and mobile. Check that the product has a comparable visual size beside neighboring items and that the theme's crop preserves useful information. Then inspect the enlarged view for artifacts that were hard to see in the thumbnail.
A consistent aspect ratio helps organize a collection, but the right composition depends on the product. A square crop that clips a long necklace needs a different framing, not a smaller necklace invented by the generator.
Repeat an approved direction across the catalog
An approved example should record more than the prompt. Keep its source product, studio choice, aspect ratio and intended gallery position together so the next person can repeat the visual direction. Describe what the frame must communicate: the item's outline, how it is worn, or its scale in a room.
Keep product accuracy separate from styling
Use the same review questions for each item, but adapt the details to the category. Skincare needs legible packaging and the correct shade. Jewelry needs the right setting and clasp. Furniture needs the correct construction and material; a generated room does not replace measurements in the listing.
Reject outputs that improve the styling by changing the product. An extra seam, a missing handle or a different surface finish can make a polished image misleading. The reviewer should have the source photos beside the output, rather than relying on memory of what the item looks like.
Publish in reviewable groups
Start with a small group of related products and inspect their storefront presentation before expanding the update. Track which result belongs to which product and variant, and keep the previous approved media available while the new set is reviewed.
If the team uses bulk generation in the full workspace, treat that as a separate creative workflow with its own review and delivery steps. The embedded app's documented add-back action applies to a selected result and product; it is not a promise that every generated batch will publish automatically.
The useful endpoint is an accurate gallery that displays correctly in the store. Keep the source product as the reference, reuse the approved creative direction, and check each result where customers will see it.
References
