Shopify clothing product image requirements (2026)

Shopify accepts product images up to 5000 x 5000 pixels, 25 megapixels, and 20 MB. Aim for 2048 x 2048 px, square, under about 200 KB. Zoom needs at least 800 x 800 px. A product can hold 250 media items. Those are the hard limits. The rest of this guide is about what to actually produce for clothing.

The numbers, in one place

The megapixel and file size caps are enforced separately, and you can fail one while passing the other. A 6000 x 5000 px export is 30 megapixels and gets rejected even if the file is only 6 MB. A 3000 x 3000 px TIFF is 9 megapixels and might still be rejected for weighing 40 MB. If an upload fails, check both before you start compressing blindly.

Why half the internet says 20 megapixels

Search this topic and you get two different answers. Some guides say 25 megapixels. Others say 20, usually alongside a recommendation to cap exports at 4472 x 4472 px.

The likely explanation is that Shopify's uploaders do not all share a cap. Product and collection media is documented at 25 megapixels. Reports from merchants suggest the Files uploader, theme assets, and blog images sit lower, around 20, which is why the same 4500 x 4500 px file can be accepted in one place and rejected in another. Animated GIFs are capped lower still.

I have not seen Shopify document the lower figure directly, so treat it as a working theory rather than a fact. But it explains the contradiction, and it explains the error most merchants actually hit, which is uploading a file to the Files section and being told it exceeds the maximum resolution after the same file went into a product page without complaint.

For apparel you are nowhere near either ceiling anyway. A 2048 x 2048 px export is 4.2 megapixels.

Shopify says PNG. For clothing, mostly use JPEG

Shopify's own documentation names PNG as the best file type for product images, with JPEG second. For apparel photography I would flip that, and it is worth saying why rather than just asserting it.

PNG is lossless, which is the right choice for flat graphics, logos, and anything needing transparency. Photographs of fabric are the opposite case. A knit texture or a printed pattern is millions of slightly different pixel values, which is exactly the content PNG compresses badly. The same garment shot exported as a PNG can be five or six times the file size of a visually identical JPEG. On a collection page loading forty of them, that is the difference between a fast store and a slow one.

Use JPEG for photographic garment images. Use PNG when you have a genuine cutout with transparency, which for most apparel sellers means the odd graphic asset rather than the product gallery.

WebP is fine to upload and generally smaller again, but there is little reason to bother. Shopify's imagery service already converts and serves whichever format the visitor's browser handles best, so uploading JPEG and uploading WebP mostly end up at the same place.

Dimensions worth producing

2048 x 2048 px square is the working standard, and it is the one Shopify recommends. It gives enough detail for zoom on a retina screen, it sits well inside every cap, and it keeps the catalogue uniform.

Square is the safer default for clothing because it holds a stable collection grid and leaves room for front, back, and detail views to sit together without odd cropping. It suits flat lays, folded knitwear, and ghost mannequin imagery particularly well, since those images are garment-shaped rather than person-shaped.

Portrait at 4:5 earns its place when fit is the thing being sold. Dresses, outerwear, denim, and anything oversized are easier to read on a model in a vertical frame, because the proportion of garment to body is the information the shopper wants. Plenty of brands run both: square for the grid, portrait inside the gallery.

What does not work is producing one image and hoping it survives every placement. Shoot or generate vertical lifestyle imagery, let the theme square-crop it for the collection card, and you get a model cut off at the shoulders with the garment pushed out of frame. The grid stops reading as a collection. If you need both ratios, export both.

Your theme decides the crop

Shopify does not display your image. Your theme does, and themes differ. Some crop collection cards to square regardless of what you uploaded. Some pad with whitespace. Some favour portrait. Some show a large hero with thumbnails beneath, others a swipeable carousel on mobile.

This is worth ten minutes before you produce a season's worth of assets. Upload a full set for one style, then look at it on a phone and on a desktop. Check how the card crops on the collection page, where the hero sits relative to the fold, and whether your detail shots are still legible at the size the theme actually renders them.

Mixed aspect ratios inside one listing are the most common version of this problem. Shopify will not block the upload, so nothing warns you. The gallery just looks uneven, and the collection grid looks worse.

What a clothing gallery has to prove

A shopper cannot feel the fabric or try the size. Every objection that a fitting room would resolve has to be resolved by the images.

Start with a lead image that makes the garment recognisable at thumbnail size. Front-facing, accurate colour, full product in frame, clean background. For a model-led brand that is a worn front shot. For a basics label or a resale operation a centred ghost mannequin or flat lay usually reads more clearly, because there is no styling to parse.

Then add the proof. Front and back, a three-quarter or side angle, at least one close-up of fabric and construction, and something that communicates fit. Photograph the collar, the cuff, the waistband, the print placement, the hardware, the lining. The test is simple: if a detail could cause a return, it needs an image. Generating front, back, and side views from one source photo is what multi-angle tools are for, and it is usually faster than rehanging the garment.

Colour has to hold across the whole gallery. A black tee should be black in all six images while keeping enough shadow detail to show the neckline and the drape. A red dress should not drift orange between image one and image four. Inconsistent colour is read as unreliability, and it produces both abandoned carts and returns. If your gallery mixes shots from different days or different sources, a retouching pass that locks white balance and exposure across the set is worth more than another angle.

Keep the campaign imagery further down the page. Heavy grading and coloured lighting make a garment look expensive and make it harder to buy, because the customer cannot tell what colour it actually is. The product gallery is a selling tool before it is a brand asset.

If any of those images are AI generated, the same rule applies with more force. A generated image that invents a pocket, smooths away ribbing, or shortens a hem is a return waiting to happen, and preserving the real garment is the whole job. We have written separately about producing on-model imagery this way. There is also a disclosure question if you sell into the EU, covered here, and it turns on whether the image counts as a deep fake rather than on whether AI touched it. For the separate question of whether you have the right to use an AI-generated image commercially, see our guide to commercial rights for AI images.

Compression fails on fabric first

There is no universal kilobyte target, because a white tee on white and a printed silk dress compress completely differently. Around 200 KB is the number Shopify points at, and for a lot of apparel that is achievable without visible loss.

What matters more is where you look when checking. Compression artifacts show up in fine detail before anywhere else, so zoom into the stripes, the embroidery, the mesh panel, the knit texture, and the dark fabric. Dark navy and black are the worst offenders and the ones most likely to go blotchy while the rest of the image still looks fine. If a garment goes soft or blocky in those areas, push quality up. If it looks identical at half the size, keep the smaller file.

Shopify will serve your images efficiently, but it cannot rescue a weak original. Blurred seams and crushed blacks stay blurred and crushed.

Variants, and the beige problem

Assign media to each variant. When a shopper picks olive, the gallery should show olive. This is basic and it is also one of the most common things left undone in a clothing catalogue, usually because the variant was added in a rush during a restock.

It matters most in the colour families nobody can distinguish by name. Cream, ivory, bone, stone, oat, sand, and beige are seven words for a range a customer cannot resolve from a swatch label, and getting the wrong one is a guaranteed return. Every one of those variants needs its own lead image plus at least one supporting shot showing the colour against something neutral.

For restocks, write the production brief once rather than deciding each time: source photo type, output ratio, required views, background standard, colour check, filename pattern. One garment photo should produce a complete set of usable outputs, not a folder that still needs sorting.

File names and alt text

Name files so a person can find them in six months. Include the SKU, the colour, and the view, such as LW102-black-front.jpg or LW102-black-detail-cuff.jpg. This pays off during reorders and variant management, not during the upload.

Alt text has two jobs. It describes the image for anyone using a screen reader, and it gives search engines something to read. "Women's black ribbed mock-neck top, front view" does both. "Image" does neither, and a string of keywords is worse than nothing. When five images show the same garment, describe what differs rather than repeating one phrase five times.

Before you publish

Open the live product page on your phone.

Can you name the colour without reading the description? Can you tell how it fits? Does zoom show you the fabric or a smear? Does one image break the pattern the others establish, through a different crop or background or edit?

If the answer to the last one is yes, that image is doing more harm than the extra angle is worth. Take it out.