AI product photos for print-on-demand mockups: t-shirts, mugs and posters

A print-on-demand buyer never touches the product before ordering, so the mockup is the product in their mind. A good mockup answers three questions in seconds: what is printed, how big is it on the item, and what will it look like in a real room or on a real person. A bad mockup answers none of them, because the design is stretched, the colours are invented, or the shirt in the photo is not the shirt the provider ships. AI can make the second and third kinds of images fast and cheap, but it cannot fix a wrong print file. This article keeps the two jobs separate, gives you typical sizes for t-shirts, mugs and posters, four prompts you can copy, and the checks that stop a pretty mockup from becoming a return.

First, two different files: the print file and the mockup image

The print file is what the printer uses. It must be high resolution, at the exact print area for that product and provider, usually as a PNG with a transparent background for apparel. The mockup image is what the shopper sees. It is a screen image, often 2000 to 3000 pixels on the long side, showing the design placed on a shirt, wrapped around a mug, or hung on a wall. Confusing the two is the most common beginner mistake. A beautiful 1080 pixel lifestyle image is useless as a print file, and a 4500 pixel print file pasted flat onto a photo, without following the fabric folds or the mug curve, looks fake even when the design itself is good.

The workflow that avoids both mistakes is simple. Design once at print size, keep that master file untouched, then make mockups from it. Never upscale a small design to print size and hope for the best, and never let a mockup tool silently stretch your design to fill a space it was not drawn for. If your design is square and the print area is portrait, redesign the layout for that area instead of stretching it. Shoppers forgive a plain design. They do not forgive a distorted one, because distortion is the first sign that the finished product will not match the photo.

AI-generated product-style image with soft light, shown as a mockup example — **AI-generated sample, made with Movapix.** Judge a mockup the way a buyer does: is the design sharp, does it sit on the product the way print would, and could you tell the colour and size from this image alone?

The sizes that actually print: t-shirts, mugs and posters

Providers differ, and that sentence matters more than any single number below. A t-shirt print area on one provider is not identical to the same shirt on another, and a mug wrap changes with the mug size. Always open your exact product in your provider's product creator and read its current template before you upload. The table below gives the typical starting points that 2026 provider guides repeat, so you can design in the right shape and check, not guess, the final numbers.

Product	Typical print file	What to check in the provider template
T-shirt, front print	4500 by 5400 pixels, PNG with transparent background, 300 dpi ideal, 150 dpi minimum for most direct-to-garment printing	Exact print area, safe margin, and whether the provider trims the edges
Mug, 11 ounce wrap	About 2475 by 1155 pixels for a full wrap, or about 2700 by 1200 pixels on some providers, PNG, 300 dpi	Handle zone, seam position, and how much of the wrap is visible from the front
Poster, 18 by 24 inches	5400 by 7200 pixels, JPG or PNG, 300 dpi, full bleed where the provider offers it	Bleed and trim, border or no border, and paper finish, because matte and glossy read differently
Listing mockup image	2000 to 3000 pixels on the long side, JPG, sRGB colour	Marketplace minimum size and aspect ratio, so the image is not cropped on mobile

Two rules follow from the table. First, resolution is pixels at print size, not a number written into the file. A small image relabelled as 300 dpi is still a small image. Second, upscale only the mockup, never as a substitute for a proper print file. Our print upscaling guide explains which sizes 2x and 4x actually reach, and the remove-background guide shows how to get a clean transparent design for apparel. If an AI image is too small for print, regenerate it larger or redesign it, rather than stretching it to fit.

The four prompts: flat, worn, wrapped and hung

Use one design description in every prompt, and change only the scene. Generate four variants per prompt and keep the one where the design is sharp, undistorted and in the right place. A mockup where the design bends correctly around a mug, or folds naturally on a shirt, beats a sharper image where the design floats on top of the fabric.

1. T-shirt flat lay, the listing hero

Photorealistic flat lay photograph of a plain black cotton t-shirt laid flat on a light wooden table, a simple white mountain line design printed centred on the chest, soft even daylight, design sharp and undistorted, no model, no wrinkles hiding the print

2. T-shirt worn, the scale scene

Photorealistic photograph of a person wearing a plain black cotton t-shirt with a simple white mountain line design centred on the chest, standing in a bright room, natural daylight, true proportions, design clearly readable, face cropped out of frame

3. Mug wrap, the front-and-curve scene

Photorealistic photograph of a white ceramic mug with a simple botanical line design wrapped around it, photographed at a three-quarter angle on a kitchen table beside a coffee pot, soft morning light, design following the curve of the mug, handle visible on the right

4. Poster on a wall, the in-room scene

Photorealistic photograph of a framed poster with a simple botanical line design hanging on a light wall above a wooden sideboard, a plant and a chair beside it for scale, soft daylight, poster straight and in proportion, no glare on the glass

For a real listing, do not stop at the hero. The set that converts is: flat or provider mockup first, worn or in-room scene second, close-up of the print third, and back or detail view fourth. The close-up matters most for posters, where a buyer wants to see line quality, and for mugs, where the buyer wants to see how the design meets the handle. If the AI close-up invents texture that the print will not have, use the flat design file itself for that slot instead. For apparel, a model-free shot keeps the fit honest: the ghost mannequin guide shows how to present a garment with no visible model. The flat prompt in that set deserves its own pass: the flat lay guide covers grid, breathing room and focal point for apparel and posters.

What a mockup must never fake

AI mockups tempt you to show a better product than you sell, and every version of that temptation ends as a return or a bad review. Do not show a colour, fabric, fit or finish you do not offer. Do not show a design larger on the shirt than the provider prints it, or a poster framed when you sell it unframed, unless the listing says so in plain words. Do not put a design on a premium blank in the photo and ship a basic blank, and do not use a model photo that implies a fit your size chart does not support.

Write the listing text to match the mockup, not the other way round. State the blank brand and weight if you know it, state whether the poster is framed or unframed, state the mug size in ounces, and state that colours can vary slightly between screens and print. That last line is not a disclaimer to hide behind. It is the honest sentence that lets a buyer who loves the design still buy it when the shade is a touch different. The same honesty rule runs through our 10 Shopify scene prompts and our small-shop photo guide, because a mockup that overpromises costs more in returns than it earns in clicks.

Cost is the easy part. In Movapix an image costs about 0.8 cents per successful generation, with no subscription, so a four-scene mockup set at four variants each costs under 15 cents in credits. The expensive part is the hour you spend checking each keeper against the real print file, the provider template and the listing text. Order a sample of your best seller as soon as it can pay for itself, photograph that sample, and replace the AI hero with the real thing. AI mockups launch a product. A sample photo keeps it selling.

Honest limits

An AI mockup must show the exact design, size, placement and product you ship. If a generated variant stretches, recolours or moves the design, regenerate it or use the provider mockup; do not publish an almost-right image and hope the buyer does not notice.
Provider templates change. The sizes in this article are typical 2026 starting points, not a substitute for the template of the exact product you sell. Check the print area, handle zone, bleed and trim in your provider account before every upload.
In Movapix, images cost about 0.8 cents per successful generation, with no subscription. Credits never expire, and an unused wallet balance can be refunded within 7 days. Commercial use is allowed, and images are private by default.
Questions first? Email [email protected] — a person reads it. For free starter credits, use @AIshopvn_bot on Telegram.

Your first images are free

Start with Google and get 40 free images + 10 videos — no Telegram, no card. Generate one mockup scene and compare it with your print file.

Start creating free