Updated November 2026 · Written and maintained by the Progression Agency strategy team
AR for ecommerce puts the product where the shopper’s doubt sits: a frame on their face, a watch on their wrist, a sneaker on their foot, a floor lamp in the corner it was meant for. As of November 2026 most of what a store needs is already built into the phone: Apple’s AR Quick Look opens USDZ models from Safari, Google’s Scene Viewer opens GLB models on Android, and Shopify serves both file types from a product page. Progression’s own senior-led team in New York plans, models, builds and measures try-on, room placement and 3D viewers for online and physical stores across the US.
On this page · 14 sections
- AR for ecommerce, seen from a single product page
- What augmented reality adds to an online store and a physical one
- Virtual try-on by face, wrist, foot and body
- View in AR: placing products in the shopper’s room
- 3D product viewers that come before the AR button
- Shopify AR and other store platforms
- AR mirrors and magic mirror technology in the store
- AI virtual try-on versus AR try-on
- Virtual try-on tools and software: build options for a store
- Getting product models ready for AR
- Measuring AR on a store: try-on to cart and returns
- What AR for ecommerce costs (market ranges only)
- Shopper privacy, camera data and adult audiences
- How Progression’s team builds AR for a store
Here for AR campaigns or a full AR app?Lenses, effects and AR ad campaigns sit on our augmented reality marketing page, and AR inside an app you own sits with augmented reality app development.
Anything labeled “Illustrative example.” describes a hypothetical store, not a real business. Platform capabilities carry the date they were true, because Apple, Google, Shopify and Snap revise their AR features every year.
AR for ecommerce, seen from a single product page
Start with one hypothetical product page. A store sells glasses and a matching desk lamp. The page has photos, but buyers keep asking whether the frame is too wide and whether the lamp is too tall for a desk. The answer the page needs is not another photo; it is the shopper’s own face and the shopper’s own desk.
Three additions change that page. A 3D viewer lets anyone turn the frame and look at the hinge. A try-on button puts the frame on the shopper’s face through the front camera. A placement button hands the lamp model to the phone’s built-in viewer so it stands on their real desk at real height. None of the three needs an app download, and each one answers a question the photos left open.
- Photos sell the look; AR answers size, fit and scale.
- The try-on button belongs next to size and color choices, not under the reviews.
- Room placement uses the viewer already built into the shopper’s phone.
- A fit note with real measurements backs up what the camera shows.
What augmented reality adds to an online store and a physical one
AR adds a step between looking and buying in which shoppers test the product on themselves or in their space. Online, that step replaces the fitting room and the tape measure. In a physical store it extends the shelf: a mirror can show every color of a frame the store does not stock, and a tablet can place a sofa that only exists in the warehouse.
That is why AR for retail is no longer only a campaign idea. A model made once for the product page also serves the store mirror, the app, paid social and the catalog, so the cost is spread across every place the product is sold. Stores that already invest in product photography find that a 3D model is the next asset in the same library.
Which products gain the most from AR?
Products where shoppers hesitate over how something will look on them or fit their space gain the most. Eyewear, makeup, watches, jewelry, shoes and furniture lead because size and appearance drive both the purchase and the return. Products bought on specification alone, such as cables or groceries, gain little.
- Face: frames, sunglasses, lipstick, foundation, hats.
- Wrist and hand: watches, bracelets, rings, nail polish.
- Foot: sneakers, boots, sandals.
- Room: sofas, chairs, lamps, rugs, wall art, appliances.
- Body: jackets, dresses and bags, where drape matters less than shape.
Is virtual try-on augmented reality?
Most of it is. A camera try-on that tracks the face, wrist or foot and draws the product over the live video is augmented reality. A newer kind of try-on uses generative AI to produce a still image of the shopper wearing the item, which is not AR at all, even though stores often label both the same way. The two are compared further down this page.
Virtual try-on by face, wrist, foot and body
Every try-on starts with what the camera has to find. A face, a wrist, a foot and a whole body each need a different kind of tracking, and the choice decides which products a store can offer, which devices will run it and how much modeling each item needs.
| What the camera tracks | Products it suits | What the shopper sees | Typical home |
|---|---|---|---|
| Face | Glasses, sunglasses, makeup, hats, earrings | The frame or shade sitting on their own face as they turn | Product page, app, social camera |
| Wrist and hand | Watches, bracelets, rings, nail color | The piece on their wrist or finger at true size | App or product page |
| Foot | Sneakers, boots, sandals | A shoe on their foot, viewed from above or in a mirror | App |
| Body | Jackets, tops, dresses, bags | The garment laid over their body in the camera | App, social camera, store mirror |
| Room surfaces | Furniture, lighting, rugs, art, appliances | The product at real size on their floor or wall | Product page, through View in AR |
| A printed image | Packaging, catalogs, labels | A 3D product or demo rising from the page | Browser, opened from a code |
How does virtual try-on find a face?
It maps the face as a mesh and pins the product to it. Google’s Augmented Faces documentation describes a center pose behind the nose, three region poses on the forehead and nose tip, and a 468 point 3D face mesh, and it names beauty and accessory try-ons as a use case. That mesh is what lets a pair of virtual glasses sit behind the bridge of the nose instead of floating in front of it. As of November 2026 Google publishes guides for Android, Unity and iOS, so a single face try-on can reach both major phone platforms.
Watches, rings and bracelets: wrist and hand try-on
Wrist and hand tracking is where virtual try-on jewelry lives. The camera finds the wrist or fingers, estimates their width and angle, and places the watch case or ring at true size, so a 44 mm case reads differently from a 38 mm one. Metals and stones need careful materials, because shoppers judge jewelry by how it catches light as the hand turns.
- Model each case size, not just each style.
- Match metal finishes to the photography so both agree.
- Show the strap or band options as variants of one model.
- Keep a fit guide with real millimeter sizes beside the button.
Sneakers and boots: foot tracking
Foot try-on runs from the rear camera pointed down at the shopper’s own feet. It suits sneakers and boots, where color and silhouette are the questions. It cannot measure a foot for size the way a fitting tool can, so the size chart stays on the page beside it.
Jackets, tops and bags: body and clothing try-on
Body tracking lays a garment over the person in the camera. Snap’s Camera Kit post of May 5, 2026 describes tracking for the face, body, hands and world, lists e-commerce virtual try-ons as a use case, and gives an example of a fashion brand running a clothing try-on Lens inside its own shopping app. Live body try-on shows shape and color well; soft fabric folding on a moving body is where the AI approach below does better.
View in AR: placing products in the shopper’s room
Room placement is the easiest AR for a store to add, because both major phone platforms ship a viewer for it. The shopper taps a button, the camera opens, and the product stands on the floor or hangs on the wall at its real size.
| Viewer | Devices | File it opens | How a store links to it | Dated note |
|---|---|---|---|---|
| Apple AR Quick Look | iPhone, iPad, Apple Vision Pro | USDZ | A link to the file, or a badge on a product image | Needs iOS 12, iPadOS 12 or visionOS 1 or later, as of November 2026 |
| Google Scene Viewer | ARCore supported Android phones | glTF or GLB | A formatted link or the model-viewer component | Android 7.0 or later, as of November 2026 |
| model-viewer in the browser | Current desktop and mobile browsers | GLB, with USDZ made for iOS | One web component on the product page | Hands off to WebXR, Scene Viewer or Quick Look |
| Shopify product media | Any device that opens the store | GLB and USDZ, up to 500 MB | Uploaded as product media | Uses Google’s model viewer and the Shopify-XR library |
| A brand app | iOS and Android | Native scenes or Lens Studio Lenses | In-app screens and links | Snap’s Camera Kit also runs on the web |
On Apple devices, the AR Quick Look gallery says Safari, Messages, Mail, News and Notes use Quick Look to show USDZ files in 3D on iPhone, iPad and Apple Vision Pro, with optional audio, Apple Pay and custom action banners. On Android, Google’s Scene Viewer guide says most Android browsers support the viewer, which also powers 3D and AR results in Google Search, and that it can add a button linking to an external page, which is how a store returns shoppers to the cart.
How does a shopper open AR on iPhone and Android?
On iPhone, a link to a USDZ file opens Quick Look. WebKit’s post on AR in Safari explains that a link marked with the rel attribute value ar and wrapping a single product image opens the live 3D view in place, shows a small cube badge on the image, and returns the shopper to the page when they close it. On Android, a formatted link opens Scene Viewer, which places the model through a button labeled View in your space, and falls back to a 3D view on phones without AR support.
- iPhone and iPad: USDZ file, Quick Look, cube badge on the image.
- Android: GLB file, Scene Viewer, a View in your space button.
- Desktop: a 3D viewer only, with a code to carry the shopper to their phone.
- Every platform: a fallback image when AR is not available.
What is the difference between USDZ and GLB?
They are two packages for the same model. GLB is the binary form of glTF, a royalty free format the Khronos consortium built for loading 3D scenes quickly in apps and on the web, and it is what Android and most web viewers read. USDZ is Apple’s choice, built on Pixar’s Universal Scene Description; the usdz specification defines it as an uncompressed, unencrypted zip archive so a phone can read it directly. A store keeps one master model and exports both.
Selling something shoppers want to see on themselves?Tell us the products and where most returns come from. The free proposal sets out the try-on type, the 3D work and how the result will be read.
Get a free proposalTell us about a project you’d like help with.
3D product viewers that come before the AR button
A 3D viewer is the version of AR that works everywhere, including laptops. It sits in the product gallery, lets the shopper turn the item and zoom into details, and on a capable phone hands the same model to AR. Google’s model-viewer web component is the common route: it runs on the last two major versions of current desktop and mobile browsers plus the last two versions of Safari, as of November 2026.
The model-viewer AR example shows the handoff in practice: one tag tries WebXR, Scene Viewer or Quick Look depending on the device, can generate the USDZ for iPhone automatically when no separate file is supplied, and can place a model on a wall instead of the floor. WebXR mode needs the page served over HTTPS. Our 3D website design team builds viewers like this into stores, and our 3D rendering services team makes the GLB and USDZ files they load.
When does a configurator beat a plain viewer?
When the product is built to order. A viewer shows one finished item; a configurator lets the shopper change fabric, size or parts, updates the price as they go and passes a buildable order to the business. Sofas, bikes, cabinets and made to order jewelry usually need the second. Our ecommerce product configurator page covers that build, and the same models feed the AR button at the end of it.
- One finished product with a few colors: a 3D viewer with AR.
- Many options that change the price: a configurator.
- Options that change the shape: a configurator with rules about what can be built.
- A product only sold in store: a viewer on the site and a mirror on the floor.
Shopify AR and other store platforms
Shopify supports 3D and AR on product pages as standard media. Its developer guide to product media lists images, 3D models, videos and YouTube or Vimeo videos as media a merchant can add, says Shopify hosted 3D models display through Google’s model viewer component, and says the Shopify-XR library opens AR Quick Look in iOS Safari and Scene Viewer on Android. Its product media API guide accepts GLB and USDZ models up to 500 MB and notes that Shopify serves models in both formats, GLB for the web and USDZ for AR on iOS, as of November 2026.
So for room placement on Shopify, the work is mostly the model and the theme: a clean, light file uploaded to each product, and a theme that shows the viewer in the gallery and puts the AR button where shoppers decide. Our Shopify app development and ecommerce website design teams handle the theme side, and Shopify SEO keeps the pages that carry the models easy to find.
How does try-on work on a Shopify store?
Face, wrist and foot try-on go beyond Shopify’s built-in media, so they are added as a custom theme section or an app block. The section opens a browser try-on scene or a Camera Kit Lens on the product page, passes the chosen variant in, and sends the shopper back to the cart with that variant selected.
WooCommerce, BigCommerce and headless stores
The same files and the same model-viewer component work on any platform that can place a web component on a product page. On WooCommerce the viewer usually goes into the product gallery through a plugin or a custom block; on BigCommerce and headless builds it goes into the product component directly. The comparison on our BigCommerce vs Shopify page covers the wider platform choice.
AR mirrors and magic mirror technology in the store
An AR mirror is a large upright screen with a camera that shows shoppers to themselves wearing products they have not picked up. Magic mirror technology is the older retail name for the same idea. Snap said at Lens Fest 2025, on October 16, 2025, that Camera Kit carries its AR to apps, websites and AR mirrors, reaching more than 68 million monthly active users in Q3 2025, and that the SDK stays free with its mandatory branding removed.
Where does a mirror earn its floor space?
In stores with more variants than shelf space, and in pop-ups that carry only a sample of the range. A mirror can show every frame color, every lipstick shade or every jacket in the line, then hand the shopper a code that opens the same product online. It also gives a quiet store something people stop for.
- Eyewear and beauty counters with deep color ranges.
- Pop-ups and launches with limited stock on hand.
- Window displays that respond to people walking past.
- Event stands where the product ships later.
Can the same try-on run on a mirror and a phone?
Yes, when it is built that way from the start. A Lens built in Lens Studio can be deployed through Camera Kit on iOS, Android and the web, and a mirror is a screen running one of those. The model, the materials and the variant list stay the same; only the camera position and the screen layout change.
AI virtual try-on versus AR try-on
AI virtual try-on generates a picture instead of tracking a camera. Google described its approach on June 14, 2023: a diffusion model takes one image of a garment and one of a person and produces a realistic image of that person wearing it, launched on women’s tops from brands including Anthropologie, LOFT, H&M and Everlane. On May 20, 2025 Google added try-on with the shopper’s own uploaded photo, and on July 24, 2025 it launched that feature in the US across Search, Google Shopping and Google Images product results, using a full length photo.
| Question a store owner asks | AR try-on | AI virtual try-on |
|---|---|---|
| What does the shopper do? | Points a live camera at their face, wrist, foot or room | Uploads or chooses a still photo |
| What does the store need? | A 3D model of each product, at real size | Clean product images on a plain background |
| What does it show best? | Scale, angle and how the piece moves | How fabric drapes and folds on a body |
| Where does it run? | Product pages, apps, social cameras, store mirrors | Search results, product pages, apps |
| What is its weak point? | Soft clothing that should fold and stretch | Exact size, and anything the photo cannot show |
When does an AI try-on fit better?
For soft clothing, where how fabric hangs matters more than exact size, and for catalogs too large to model in 3D. A generated still image can show a dress or shirt on many body types from product photos the store already has.
Can a store run both?
Many should. AR suits rigid products where scale and angle decide the sale: glasses, watches, shoes, furniture. Generated images suit apparel. A store selling both can put camera try-on on accessories and a photo try-on on clothing, and track both with the same measures.
Virtual try-on tools and software: build options for a store
Store owners have six realistic routes, and most stores combine two of them. The right mix depends on the products, the platform the store runs on and whether the brand already has an app.
| Option | How it reaches shoppers | Good fit | What to weigh |
|---|---|---|---|
| Built-in viewers (Quick Look, Scene Viewer) | A link or button on the product page | Room placement for furniture, decor and appliances | Placement only; no face or body tracking |
| Web components such as model-viewer | Inside the product page itself | A 3D viewer that opens AR on demand | Model weight and page speed |
| WebAR with WebXR or an open engine | In the phone browser, from a link or code | Try-on and packaging scenes without an app | Browser tracking is lighter than native |
| Camera Kit Lenses | Inside a brand app or website | Face, body and clothing try-on built in Lens Studio | Access by application and Snap’s approval |
| A native ARKit and ARCore build | A brand’s own iOS and Android app | Precise wrist, foot and hand try-on | App store releases and upkeep |
| AI image try-on | A photo upload on the product page | Apparel where drape matters | Needs clean garment images; still frame only |
Does a store need its own virtual try-on app?
Only when the try-on must be precise enough that a browser cannot deliver it, or when the brand already has an app shoppers use. Wrist and foot try-on at true scale, and try-on across a large catalog, are the usual reasons. For most stores a browser try-on on the product page reaches more people, because nothing has to be installed. Our iOS and Android teams build the app route when it is the right one.
What happened to Instagram try-on effects and 8th Wall projects?
Two tools stores once relied on have changed. Meta’s Meta Spark update of August 27, 2024 ended third-party Meta Spark tools and content effective January 14, 2025, which took brand-built try-on effects out of Instagram and Facebook. The 8th Wall migration FAQ set the end of platform access at February 28, 2026; hosted projects stay live but locked until February 28, 2027, after which hosting ends. A store with a try-on still hosted there should move it to an open engine or its own site before that date.
The browser route keeps getting stronger. The W3C’s WebXR Device API, which lets web pages use a phone’s AR, was updated as a Candidate Recommendation Draft on November 9, 2026. Social try-on continues on Snapchat and TikTok, covered on our branded AR filters page.
Have CAD files or a product photo library?Send a sample of what exists. The free proposal covers the models, the AR button on each product page and the platforms it must run on.
Get a free proposalTell us about a project you’d like help with.
Getting product models ready for AR
Every AR feature runs on a 3D model, and the model is where most of the time goes. A store either has CAD files from manufacturing, has physical samples that can be scanned or modeled, or has only photos.
Apple’s augmented reality page says RealityKit’s Object Capture can create 3D objects optimized for AR in minutes using photogrammetry, which suits rigid products with matte surfaces. Glossy, transparent and very thin items, such as glassware, lenses and fine chains, still need modeling by hand. A product family that shares a shape, such as one frame in eight colors, needs one model and eight materials.
How accurate does a try-on model need to be?
Accurate to the millimeter on the dimensions shoppers judge, and light enough to open quickly on a mid range phone. A watch case or a frame width that is slightly off undermines trust in the whole store, while hidden detail inside a product adds weight and nothing a shopper will see.
- Real world scale checked against the product’s spec sheet.
- Materials matched to the product photos under neutral light.
- File size kept small enough to open on mobile data.
- A named owner who updates models when products change.
Measuring AR on a store: try-on to cart and returns
A store measures AR the way it measures any product page feature: by comparing shoppers who used it with shoppers who did not. The comparison has to be set up before launch, with every AR button and try-on event tagged, or the results will be guesses.
| Measure | Question it answers | How to collect it |
|---|---|---|
| AR opens per product page view | Do shoppers notice the button? | A tagged click on the AR button |
| Time in the viewer or try-on | Did it hold their attention? | Viewer events sent to analytics |
| Add to cart after AR, against shoppers who skipped it | Did it settle the buying doubt? | Store analytics, split by AR use |
| Returns by reason, before and after | Did fit and look returns change? | Return reasons in the order system |
| Variants tried per session | Are shoppers exploring the range? | Try-on events by variant |
| Devices that failed to open AR | Who is being left out? | Error and fallback events |
How should returns be compared?
By reason and by product, over a long enough period to cover a full return window. If a frame’s returns for “too wide” or a sofa’s returns for “too big” fall after AR launches on those products, while returns for other reasons hold steady, the AR is doing its job. Our conversion rate optimization team runs these comparisons as controlled tests where traffic allows.
What AR for ecommerce costs (market ranges only)
Cost depends on how many products are modeled, how precise the try-on must be and how many places it runs. The bands below are published US price bands (market ranges only), typical low to published high; Progression quotes each project in a written proposal after scoping.
| Store AR work | Typical low to published high |
|---|---|
| 3D product viewer with AR added to a store already running | $8,000 to $40,000 per project |
| Product models prepared from CAD files | $2,000 to $20,000 per product family |
| AR placement or try-on added inside an existing app | $30,000 to $80,000 per project |
| Standalone product visualization app | $60,000 to $150,000 per project |
| 3D product configurator with live pricing | $30,000 to $200,000 per project |
| Hosting, delivery and upkeep for 3D files | $300 to $3,000 per month |
What drives the cost of a try-on build?
Face, wrist, foot and body try-on carry no single market band, because the scope varies so widely. Five things set the price: the tracking type, the number of products and variants to model, the platforms it must run on, whether it lives in a browser, an app or a store mirror, and how deeply it connects to the cart and the catalog. A free proposal turns those into a number for a specific range of products.
- Tracking type: face, wrist, hand, foot, body or room.
- Products and variants: ten frames in four colors is forty looks.
- Surfaces: product page, app, social camera, store mirror.
- Catalog link: manual uploads or an automated product feed.
- Upkeep: new models each season as the range changes.
Pricing depends on scope and is confirmed in a written proposal. Published prices are subject to change.
Shopper privacy, camera data and adult audiences
Progression plans store AR for adult shoppers, 18 and over, the same audience line our Gen Z marketing agency work follows. A try-on asks for the camera, so the request explains why before the browser asks, and the experience works on the device without keeping images of the shopper’s face.
- Camera permission asked only after the shopper taps try-on.
- No storage of face or body images beyond what the platform itself handles.
- Photo uploads for AI try-on deleted on a stated schedule.
- Plain wording wherever an email or phone number is collected.
How Progression’s team builds AR for a store
One senior-led team takes a store from the first brief to the measured result: product selection, 3D modeling, try-on and viewer builds, theme and app work, mirror setups, analytics and the report. The models stay with the store and are reused in ads, catalogs and marketplaces.
- Pick the first products from sales and return reasons.
- Collect CAD files, samples or photos for each one.
- Model, check scale and export GLB and USDZ.
- Add the viewer, try-on and AR button to the product pages.
- Tag every AR event and compare against shoppers who skipped it.
- Extend to more products, the app or a store mirror once the first set proves out.
Related work sits on these pages:
- Augmented reality app development for AR inside a brand app.
- AR and VR app development for headset and visualization apps.
- 3D rendering services for GLB and USDZ product models.
- Ecommerce product configurators for made to order products.
- Augmented reality marketing and branded AR filters for campaigns.
- Metaverse marketing and metaverse app development for 3D worlds.
- Ecommerce marketing and Shopify marketing for the traffic that reaches the AR.
- Retail marketing for store floors and windows.
- Fashion, beauty and jewelry marketing for the categories that lean on try-on.
- Ecommerce app development and ecommerce website redesign for the store itself.
- AEO for ecommerce and ecommerce SEO so the product pages get found.
What this page leaves out: the Progression Method, proprietary and reserved for clients. Reach out to learn more
Getting found in search
AI, AEO and what is changing
Paid media and lead generation
Websites and design
Choosing and working with an agency
Software and app development
Website design by industry and type
Web development, platforms and hosting
Social, content and brand
By industry and by situation
Frequently asked questions
What does AR for ecommerce actually do for a store?
How does a virtual try-on app know where glasses or makeup should sit?
Is augmented reality try-on only for fashion and beauty?
What happens when a shopper taps View in AR?
Which 3D files does a store need for AR?
Does Shopify support AR?
What are AR mirrors?
What is magic mirror technology?
How is AI virtual try-on different from camera try-on?
Can shoppers try clothes on using their own photo?
Do shoppers have to download an app to try products on?
Which phones can open View in AR from a store?
Can virtual try-on show rings and earrings?
How much does it cost to add AR to an online store?
What should a store track after adding try-on and room placement?
Can AR lower returns for fit and size?
Can a store still publish its own try-on effect on Instagram?
What should a store do with a try-on hosted on 8th Wall?
Who makes the 3D models for an AR store?
Does a store need separate try-on builds for its app, site and mirror?
Is AR for retail worth it for a small product range?
Selling something shoppers want to see on themselves?Tell us the products and where most returns come from. The free proposal sets out the try-on type, the 3D work and how the result will be read.
Get a free proposalTell us about a project you’d like help with.
Get a free marketing proposal
Tell us what you are trying to grow and we will come back with a plan, not a pitch deck. Same-day reply on weekdays.
