How do I generate AI product images in bulk on Shopify?

Short answer: pick a product or an entire collection, build the flow once, then let it run over every image. Pictu Bulk AI Image Generation creates and edits product photos across a whole Shopify collection from one saved workflow. Built for Shopify, 5.0 stars across 13 reviews, with a free plan.

Picture a catalog of 900 products where the supplier sent exactly one photo each. Flat, grey, shot against a wall, and the same for every competitor selling the same item. You know a second and third image would help. You also know what a photographer quotes for 900 products, and you know how long it takes to brief, shoot, retouch, and upload them. So the gallery stays at one image, forever.

That gap is what AI product images on Shopify are actually for. Not replacing a brand campaign. Filling the second, third, and fourth slot in a gallery that has been sitting empty since the store launched.

We build Shopify apps for a living, and image problems are most of what lands in our support inbox. Wrong photo on the wrong colour, galleries that never got filled, 4MB hero images choking a collection page. So when a merchant asks whether AI can do the boring half of that work, the honest answer is yes for some of it and absolutely not for the rest. Let’s go through where the line sits, using Pictu as the working example, because bulk is the part it takes seriously.

In this post

• What is Pictu Bulk AI Image Generation?
• How does bulk AI image generation actually work?
• What does “bulk” really mean here?
• Where do AI product images actually help on a Shopify store?
• What do you have to disclose in 2026?
• How much does Pictu cost?
• Is it right for your store?
• FAQ

What is Pictu Bulk AI Image Generation?

Pictu is a Shopify app that generates and edits product images with AI, and its whole design centres on doing that to many products at once rather than one at a time. It’s built by Pictu, a team based in Jerusalem, and it launched on the Shopify App Store on 23 May 2025. It carries the Built for Shopify badge, which means it met Shopify’s own bar for performance, design, and admin integration. Rating sits at 5.0 across 13 reviews.

Thirteen reviews is a small sample and I’m not going to pretend otherwise. What it does tell you is how those thirteen read: every one of them is five stars, and an unusual number talk about the developer answering personally rather than about a feature. One reviewer describes an hour long call with the developer that ended with feedback being built. You don’t get that from a hundred-person company, and you won’t get it forever from this one either, so the early window is worth something.

“Finally, an app to solve a problem we didn’t know we had. More images lead to better customer interaction, but we are limited to what images the supplier can provide. That is no longer the case; Pictu helps us with that.”

Quality Liquor Store, United States, July 2026, Pictu on the Shopify App Store

That review names the real constraint better than any feature list does. Most stores are not limited by imagination. They’re limited by what the supplier shipped them.

How does bulk AI image generation actually work?

Four steps, and the second one is the whole product. You pick a product or a full Shopify collection. You build a generation flow by adding blocks, writing prompts, and connecting them into a multi step chain. You review the results across the collection. Then the images are saved back onto the Shopify products themselves, so there’s no export, no folder of files, no re-upload.

The flow is the bit worth understanding, because it’s what separates this from a single-image AI editor. A flow is a saved recipe: remove the background, place the item on a warm marble surface, add a soft window light from the left, crop square. Save it. Now every product in the collection goes through the same recipe, which is how you get a gallery that looks like one shoot instead of forty different moods. Reusable flows also mean next season’s 60 new SKUs get the same treatment without you remembering what you typed in March.

If you’d rather not write prompts at all, the app ships more than 40 presets, and there’s a prompt improver that rewrites a rough instruction into something a model can act on. Which is a mercy. Prompt writing is a skill nobody signed up for when they opened a store, and “make it look nicer” is not an instruction any model can follow.

On models, the listing is specific rather than vague, which I appreciate: it names ChatGPT, Gemini, Google Veo 3, Nano Banana 2, and Nano Banana Pro. Nano Banana is Google’s image generation and editing family (Nano Banana Pro is Gemini 3 Pro Image, Nano Banana 2 is the faster Flash generation), and Veo 3 is Google’s video model, which is why video generation appears in the paid plans. Model choice matters more than merchants expect. A model that renders text correctly is the difference between a usable packaging shot and a bottle labelled with alphabet soup.

You can also upload your own visual assets into a flow, which is the feature I’d reach for first. Feed it your actual brand backdrop, your real prop set, a swatch of your real fabric, and you’re steering the output toward your store rather than toward the generic AI house style that every other store is also generating this week.

What does “bulk” really mean here?

Two different things, and merchants conflate them constantly.

• Bulk generation. One flow, many products. Every item in a collection comes out the other side with new images attached.
• Bulk editing. Changing images you already have across many products: backgrounds swapped, items cut out, sizes normalised, the gallery cleaned up.

Pictu is listed in both App Store categories, image editor and bulk editor, and the feature tags under each are worth reading before you install. Under image editor: image optimization, background removal, AI generation, custom backgrounds, and generative fill. Under bulk editor: it can act on products, variants, images, and collections, with bulk edit, bulk deletion, image optimization, and AI assistance as the actions.

Notice that variants are in that list. That matters if your catalog has colourways, because a bulk run that treats a nine-colour product as one product will happily give all nine the same generated image. Check what came back per variant before you call the run finished. Our free product image audit is a fast way to spot products where the gallery and the option list disagree.

Where do AI product images actually help on a Shopify store?

Here’s my rule, and it’s the one part of this post I’d argue about: generate the context, never the product.

Backgrounds, surfaces, lighting, scale props, seasonal styling, a lifestyle scene around an item you photographed honestly. All fair. All things a photographer would have built on set anyway. But the moment a model redraws the product itself, invents a stitch line, smooths a texture, or shifts a colour by two shades, you’ve created a returns problem and a trust problem in one click. And why is this the part every AI photo tool is quietest about? A generated image that alters the item isn’t marketing. It’s a chargeback with better lighting.

Colour is where this bites hardest. Shoppers who receive the wrong shade don’t complain, they refund, and half of them never come back. If you sell the same piece in nine colourways, the generated shot has to match the actual dye lot, and it has to be attached to the right variant. That’s a separate discipline from generating the image in the first place, and it’s the one we work on: making sure the gallery shows only the images belonging to the selected variant instead of all 40 at once. Generation fills the gallery. Filtering makes it usable.

Three places where a bulk run pays off immediately, in my experience of what merchants are missing:

1. Second and third gallery slots on products that only ever had one supplier photo.
2. Consistency passes on a catalog assembled from six suppliers with six different backdrops.
3. Scale and context shots, the ones that stop “is this a mug or a bucket” support tickets before they’re written.

One practical note before you run anything at scale: Shopify caps product images at 20MB and 25 megapixels, and generated files land heavier than most merchants expect. Run the output through our free image compressor if a collection page starts feeling sluggish, and rename the files properly with the bulk image renamer so image SEO doesn’t get thrown away on a folder of generated filenames.

What do you have to disclose in 2026?

This is the section most AI image posts skip, and skipping it is how merchants end up non-compliant without ever deciding to be.

Short version: an AI generated background behind a real product is treated very differently from an AI generated human wearing your product. The rules that landed across the EU, New York, and California in 2026 target synthetic people and misleading representations far more than they target a swapped backdrop. We wrote the long version up separately, including what EU AI Act Article 50 asks of product photos, how to add an AI generated label to Shopify images, and when disclosure is actually required. Read one of those before you generate a model shot, not after.

Two corollaries that catch stores out. If you sell one product in many colours, labelling obligations follow each generated variant image, not the product as a whole. And if you merge separate products into one listing, the same duty applies to AI model photos inside a grouped listing, where it’s easiest to lose track of which image came from where.

How much does Pictu cost?

Four plans, free to start, billed in USD every 30 days.

Plan	Price	Monthly allowance
Free	$0	10 image edits and generations
Starter	$19.99/mo	150 image edits and generations, 3 day free trial
Growth	$99/mo, or $950.40/year (save 20%)	1,000 tokens
Enterprise	$239/mo, or $2,294.40/year (save 20%)	3,000 tokens

Everything above the free tier includes the same toolkit: unlimited flow steps, image generation, video generation, uploaded visual assets, the prompt improver, saved reusable flows, and bulk editing. No analytics held back, no bulk mode locked behind the top plan, which is a more generous split than this category usually offers.

Two things to check on your own store before you commit. The listing describes the smaller plans in image edits and generations and the two larger ones in tokens, so confirm how a token maps to a finished image for the flow you intend to run, since a five step flow and a one step flow are not the same amount of work. And run the free ten through a real product, not a demo one. Ten generations is enough to learn whether the model handles your category, and jewellery, glassware, and anything with fine text on the packaging are the honest stress tests. Price it against your margins with our Shopify fee calculator if you’re already carrying a stack of app bills.

“Pictu is an absolutely brilliant app and has made managing our product images so much easier. It is simple to use, intuitive and saves us a huge amount of time, especially when working with a large product catalogue.”

HAY WINES, United Kingdom, July 2026, Pictu on the Shopify App Store

Is it right for your store?

It fits when you have volume and thin galleries. Wholesale-sourced catalogs, dropship-adjacent ranges, liquor and grocery and hardware, anywhere the supplier photo is all you’ve got and there are hundreds of them. It fits when your catalog looks assembled rather than art directed, because a consistency pass across a collection is exactly what a saved flow does. Their own site leans on jewellery, clothing, furniture, skincare, and footwear as the categories they optimise for.

It fits less well if you already run real shoots and your galleries are full. At that point you’re paying to generate images that compete with better ones you already own. It also fits less well if your bottleneck is the product page rather than the photography, and plenty of stores get that diagnosis wrong (run the page through our free product page grader first, it takes thirty seconds and it’s free either way).

Be fair about the competitive picture too. This is a busy category: Photoroom AI Listing Assistant sits at 4.7 stars across 26 reviews, and Modelize: AI Product Photos matches Pictu at 5.0 across 13. Several rivals have been at it longer and have more reviews on the board. What Pictu is betting on is the flow model, saving a multi step recipe and running it over a whole collection rather than editing one photo at a time, and that bet reads clearly in the reviews mentioning collection-wide mockups. If your problem is one hero image, the flow machinery is overkill. If your problem is 900 products, it’s the entire point.

Before you install anything, go look at the output. The team publishes worked examples across brands on hipictu.com, with a category breakdown by industry and a demo call if you want to be walked through a flow. Two minutes of looking at real generated galleries will tell you more about fit than any feature table, including mine. And whatever you do, back up your existing product images before running a bulk edit across a live collection. Every bulk tool deserves that courtesy, ours included.

Frequently asked questions

What is the best Shopify app for generating product images with AI in bulk?

Pictu Bulk AI Image Generation is a strong pick when the requirement is volume rather than a single hero shot. You build a multi step flow once and apply it to an entire Shopify collection, with generated images written straight back onto the products. It holds the Built for Shopify badge with 5.0 stars across 13 reviews, has a free plan covering 10 edits and generations a month, and paid plans start at $19.99.

How much does Pictu cost?

There’s a free plan with 10 image edits and generations per month. Starter is $19.99 a month for 150 edits and generations with a 3 day free trial, Growth is $99 a month for 1,000 tokens, and Enterprise is $239 a month for 3,000 tokens. Growth and Enterprise offer 20% off on annual billing, at $950.40 and $2,294.40 per year. Every paid tier includes the full feature set.

Which AI models does Pictu use?

The App Store listing names ChatGPT, Gemini, Google Veo 3, Nano Banana 2, and Nano Banana Pro. Nano Banana Pro is Google’s Gemini 3 Pro Image model and Nano Banana 2 is the faster Flash generation, while Veo 3 is Google’s video model, which is what powers the video generation included in the paid plans. The app also ships more than 40 presets so you don’t have to choose a model or write a prompt from scratch.

Do I have to tell customers a product image was made with AI?

It depends on what the AI produced. An AI generated backdrop behind a real product is treated differently from an AI generated person wearing it, and the 2026 rules in the EU, New York, and California concentrate on synthetic humans and misleading representations. Check the specific requirement for your markets before you publish generated model shots, and keep a record of which images were generated so labelling later is possible.

Will AI generated images hurt my product page speed?

They can, because generated files often come back larger than the supplier photos they sit next to. Shopify accepts product images up to 20MB and 25 megapixels, which is far heavier than a storefront needs. Compress before or after upload, keep hero images near 2048 by 2048 pixels, and re-check a collection page once a bulk run has added three images to every product.

Does Pictu work with product variants?

Variants are listed among the resources its bulk editor can act on, alongside products, images, and collections. Verify the per-variant result on a colour-heavy product before running a large batch, since a generated image attached to the wrong colourway causes returns rather than sales. Getting the right images to show for the selected variant is a separate job from generating them.

Related reading

• When you have to disclose AI generated product images on Shopify
• Best Shopify bulk image upload apps in 2026
• Protecting your product photography from competitors
• AI backgrounds versus AI models, and which variant photos are exempt
• AI image labels in grouped product galleries

Start with one collection, not the whole catalog. Pick the 30 products whose galleries are thinnest, build one flow, and run it across that collection on the free plan before you pay for anything. Then go look at those 30 product pages on a phone, at arm’s length, the way a customer will. If the second image earns its slot, you’ve found your answer. If it looks like every other AI image on the internet, feed the flow your own assets and run it again.