September 10, 2026 · 7 min read · by Aashirvad Kumar
Google's virtual try-on lets a shopper see a garment rendered on a body, either their own uploaded photo or a diverse set of AI-generated models, right inside Search and Shopping. The feature is powered by generative AI, but it does not invent your product out of thin air. It reads the exact product photo you already submit in your shopping feed, then re-renders that same garment onto a model so it drapes, stretches, and wrinkles realistically. That means the quality of the try-on a shopper sees is capped by the quality of the still image you upload. Feed a clean, full, front-facing shot and Google renders you correctly. Feed a busy lifestyle scene with props and cropped sleeves and the render degrades or your product gets skipped entirely.
This is a different surface from the one that decides which thumbnail appears in an AI Overview, which we cover in our guide to how AI Overviews pick product images. Virtual try-on is a rendering engine with strict input rules, not a ranking beauty contest. Understanding the google ai try-on product photo requirements is the difference between your listing appearing as a try-on-enabled result and it sitting flat while a competitor's coat gets virtually worn by thousands of shoppers. Below is exactly which photo Google needs, which categories qualify, and how to shoot or generate a try-on-ready image.
Key takeaways
Google needs a single, high-resolution image showing one garment on a front-facing model or mannequin in a simple pose, or the item laid flat on a surface. According to Google Merchant Center Help, the model should be facing directly forward with arms down at the side, and the frame should show the entire garment with nothing obscuring it. No hands crossed over the chest, no handbags, no accessories layered on top. The generative model has to understand the full shape, seams, and drape of the garment before it can re-render that garment onto a different body, and any occlusion forces it to guess. Meeting the google ai try-on product photo requirements starts with this one rule: isolate the product and show all of it.
This is why a hero lifestyle shot, the kind that wins on a landing page, is often the wrong input here. A model laughing on a beach with sunglasses and a tote works for emotional appeal but hides the hemline, the sleeve length, and the neckline the try-on engine depends on. Keep your emotional imagery for other placements and give try-on a clean, honest garment shot.
Aim above the floor: the minimum is 512 by 512 pixels, but Google recommends 1,024 or higher. A 1,024-plus image gives the model more fabric detail and renders noticeably cleaner, so treat 1,024 as your real minimum.
The technical floor is 512 x 512 pixels, with Google recommending 1,024 pixels or higher for the best result. Anything below the minimum simply will not qualify, and images that only scrape the floor tend to produce soft, low-detail renders. Higher resolution gives the AI more fabric texture, stitch detail, and color accuracy to work with, so a 1,024-pixel-plus image consistently renders cleaner than a barely-compliant one. Google's own Shopping blog frames the technology as accurately capturing how clothing folds and casts shadows across different body types, and that fidelity is only possible when the source pixels are there to read. Treat 1,024 pixels as your real minimum, not the 512 floor.
| Supported | Not supported |
|---|---|
| Tops | Lingerie |
| Bottoms | Bathing suits |
| Dresses | Accessories |
| Shoes | Other categories, for now |
Virtual try-on currently supports tops, bottoms, dresses, and shoes, while lingerie, bathing suits, and accessories are explicitly excluded. If you sell within a supported category, your correctly formatted product photo becomes eligible to render on Google's models, which range across sizes from XXS to XXXL so shoppers can see the fit closest to their own body. If you sell an excluded category, no amount of image optimization will surface a try-on today, so route that effort into standard feed image quality instead. It is worth re-checking Google's category list periodically, because the supported set has expanded since the feature launched and is likely to keep growing.
One nuance for tops specifically: Google asks that sleeves be rolled down, hoods positioned down, zippers closed, and the garment styled untucked. These constraints exist so the engine sees the garment in its default, fully-extended form, which is the version it needs to map onto a model. A half-zipped hoodie with the hood up gives the AI an ambiguous silhouette.
There are two try-on experiences, and they place different demands on your photo. The merchant-facing version renders your garment onto Google's diverse pool of AI models, and that render is driven entirely by the product photo in your feed. The shopper-facing version, described in Google Shopping Help, lets a person upload a single full-body photo of themselves, or in the US create a digital avatar, and see your garment on their own body. In both cases your submitted product image is the garment source of truth. The shopper controls the body; you control the clothing. That split is why nailing the google ai try-on product photo requirements matters so much: it is the one input you own on both surfaces.
Because the clean, isolated garment shot serves both experiences, you get double value from producing one correct image. The same photo that qualifies you for model-based try-on in Shopping also feeds the shopper's self-upload experience.
You can meet every requirement either with a controlled studio shoot or by generating the shot, and for most sellers a hybrid approach is fastest. Photograph the garment flat or on a front-facing mannequin against a plain background, keep the lighting even, and make sure the full item is in frame with the sleeves and hem uncropped. If your only existing asset is a cluttered lifestyle photo, this is where generative tooling helps. Our approach to AI product photography can take a single garment reference and produce a clean, front-facing, high-resolution frame that hits the pose and resolution rules without a full reshoot, and the same discipline applies whether you are optimizing for Google or building Amazon product photography. For deeper guidance on the studio fundamentals behind these shots, our clothing product photography guide walks through lighting and styling in detail.
Whatever method you use, verify the output against Google's checklist before you submit: one garment, front-facing or flat, entire item visible, nothing obscuring it, and at least 1,024 pixels. Clean product imagery is increasingly the gate to every AI shopping surface, a shift we unpack in our piece on product images as the new ranking signal. Get the input right once and you unlock try-on, cleaner feed thumbnails, and better AI Overview eligibility from a single asset.
Turn one garment reference into a front-facing, high-resolution frame that hits Google's pose and resolution rules without a full reshoot. 50 free credits, no credit card.
Start free →Usually no. Google automatically opts in brands that have a shopping feed and high-quality imagery, so if you sell a supported category your existing product photo is already being evaluated. You can opt out of the experience by contacting Google Customer Support from your Merchant Center account.
The minimum is 512 x 512 pixels, but Google recommends 1,024 pixels or higher. Higher resolution gives the AI more fabric detail to render, so treat 1,024 pixels as your practical floor rather than the 512 minimum.
It is not ideal. Google needs one garment on a front-facing model or mannequin in a simple pose with arms down, or laid flat, with nothing obscuring the item. A busy lifestyle scene with props, accessories, or a dynamic pose hides the garment details the engine needs and can disqualify the image.
Tops, bottoms, dresses, and shoes are currently supported. Lingerie, bathing suits, and accessories are excluded. The supported list has expanded over time, so it is worth re-checking Google's documentation periodically.
Both experiences exist. In Shopping, Google renders your garment onto a diverse set of AI models sized XXS to XXXL. Shoppers can also upload a single full-body photo of themselves, or in the US create a digital avatar, to see the garment on their own body. Your submitted product image is the garment source in every case.
Comments
No comments yet, be the first.
Leave a comment