Make a product photo without a photoshoot

Cuts your real product out and restages it. The product itself never gets redrawn.

Use it when

You have a photo of the actual product and the label, logo and proportions have to survive intact — which, for anything you sell, they do.

Cost

Free tier. API from $0.02/image.Free tier. Paid tiers are Pro, Max and Ultra with rising credit and export allowances — check current pricing, the site renders the numbers dynamically. The API is published flat: $0.02 per image for background removal, $0.10 per image for the generative editing endpoints (AI backgrounds, shadows, relighting).

Good at

Batch, and having somewhere to grow. Phone app, web app, Shopify integration and an API all sit on the same engine, so one Etsy listing and a 500-SKU catalog use the same tool. Cutout edge quality benchmarks at or near the top of the field.

The catch

It runs on AI credits, and the features you actually came for — AI backgrounds, shadows, relighting — burn them fast, with 4K output and adjustable shadows fenced off on the upper tiers. The generated backgrounds also have a house look: soft studio surface, tasteful shadow, slight haze. One image is fine. Twenty in a row and your catalog announces how it was made.

Wrong for

A scene that doesn't exist yet — a hand holding your product, your bottle on a mountainside at golden hour. That's generation, not compositing.

Upload the product, describe the scene, get the scene. Free, in an app you already have.

Use it when

The product needs to be somewhere a compositor can't put it — held, used, lit from inside the environment rather than pasted onto it.

Cost

Free in the Gemini app; per-image on the API.Free in the Gemini app. Per-image on the Gemini API and Vertex AI. Outputs from 512px to 4K.

Good at

Understanding a scene rather than assembling one. It relights the product to match the environment, holds multiple objects consistent across a set, and you can keep talking to it — 'lower the camera', 'warmer light' — instead of starting over.

The catch

It redraws your product. Shape and colour normally survive; small type, logo letterforms and fine label detail drift, and they drift silently — the image looks right until you zoom in on the ingredients panel and find an invented word. Every single output needs checking at full size, which is exactly what kills it as a batch tool.

Wrong for

Regulated packaging, anything where label text is a legal claim, or a run of 200 SKUs nobody is going to inspect one by one.

A canvas where you place the product, the props and the light yourself.

Use it when

You already know what the shot should look like and you're tired of trying to describe it in a sentence. You want to move the bottle two inches to the left.

Cost

Free trial, then $8–$38/mo.Free: 1 custom model, 5 images, 1 video. Pro $8/mo. Pro+ $26/mo (up to 8 custom models, 80 images, 3 videos). Scale from $38/mo, which is where API early access starts. Enterprise custom.

Good at

Art direction. Drag-and-drop scene composition, custom models trained on your specific product, and on-model / virtual try-on for apparel — the things a prompt box structurally cannot give you.

The catch

The quotas are small for the money. Eighty images a month on the $26 tier is one afternoon of iterating, and iterating is the entire reason to use a canvas tool rather than a prompt. You end up rationing the thing you bought it for. API access is gated to the $38 tier and still labelled early access.

Wrong for

One quick photo. The setup only pays back if you're building a look you'll reuse.

Free, and already sitting on top of your catalog — if your catalog is in Merchant Center.

Use it when

You sell through Google Shopping and the feed images just need to stop being bad: backgrounds swapped, resolution raised, a short animation added.

Cost

Free for Merchant Center users. No paid tier.Free for all Merchant Center users. There is no paid tier.

Good at

Being free and being where the products already are. It edits feed imagery in place — no export, no re-upload, no filename discipline — which for a catalog is most of the actual labour.

The catch

The restriction list is long and you'll hit it. It won't process images containing hands, wall art or light fixtures; it refuses regulated goods outright — alcohol, tobacco, pharmaceuticals, weapons, gambling, health and medical devices; it's unavailable in Russia and France, and video generation only runs in six countries. There's also a cap of 70 images per product. Find that out after you've moved your workflow, not before.

Wrong for

Anyone not selling through Google, and anyone who wants a distinctive look. This makes a weak catalog acceptable; it will not make a good one memorable.

For when 'a lot of images' means thousands and you'd rather call an API than click.

Use it when

Catalog imagery is a pipeline for you — enhance, cut out, restage, upscale — running on a schedule rather than by hand.

Cost

Free 50 credits, then $9–$39/mo.Free account with 50 credits. Essentials $9/mo, Professional $39/mo (includes a 200-credit API starter pack). Self-serve API from $59 for 1,000 credits. Operations consume 1 to 10+ credits each depending on complexity.

Good at

Volume plumbing. Background removal, scene generation, upscaling, on-model fashion and image-to-video behind one API, with batch treated as the normal case rather than a feature.

The catch

The credit-per-operation model makes real cost genuinely hard to forecast. An operation is 1 credit or 10-plus depending on what it is, so '1,000 credits for $59' is emphatically not 1,000 images, and you can't know your rate until you know your mix. Price a representative sample before you commit a catalog to it.

Wrong for

One product and one photo. You would be buying an API to do what a free phone app does.