AR try-on

What is AR try-on?

AR try-on is an augmented reality experience that overlays a virtual product onto a live view of the shopper or their environment, letting them preview how the item looks before buying. A camera feed from a phone, tablet, webcam, or smart glasses serves as the canvas, and a 3D model of the product is anchored to the relevant anchor point — a face for eyewear and cosmetics, a wrist for watches, a foot for sneakers, or a floor plane for furniture. The shopper can rotate their head, walk around the room, or move the camera, and the virtual item responds in real time with adjusted scale, perspective, lighting, and occlusion.

The category sits at the intersection of computer vision, 3D content production, and ecommerce. It is delivered through native mobile apps, web browsers using WebAR or WebXR, and social platforms such as Snapchat Lens Studio, TikTok Effect House, and Meta Spark. AR try-on is treated as a merchandising surface rather than a novelty, with retailers measuring its impact on conversion rate, average order value, return rate, and time on product.

How AR try-on works

An AR try-on session moves through a pipeline that combines tracking, rendering, and product data. The stages below describe the path from camera permission to a final add-to-cart event.

1. Camera and consent. The experience requests camera access through the browser or app. On the web this uses the getUserMedia API; in native apps it uses platform permissions. Without this grant the experience cannot proceed, so onboarding copy and fallback states matter.
2. Tracking and anchoring. A tracking engine identifies the relevant surface or feature. Apple ARKit and Google ARCore handle world tracking and plane detection for furniture and footwear. Face tracking meshes from ARKit, ARCore Augmented Faces, MediaPipe, or proprietary lens platforms identify landmarks for eyewear, jewelry, headwear, and makeup. Hand and wrist tracking support watches and rings. WebXR exposes a subset of these capabilities to compatible browsers.
3. Asset loading. A 3D model — typically glTF, GLB, or USDZ — is fetched from a CDN along with PBR textures, normal maps, and any rigging or morph targets. Tools such as Google model-viewer, three.js, Babylon.js, or platform-native renderers handle decoding and scene setup. Asset weight directly affects load time on mobile networks.
4. Calibration and sizing. For categories where fit matters, the experience estimates real-world dimensions. Eyewear apps measure pupillary distance and face width; footwear apps estimate foot length using a reference object or a depth sensor; furniture apps use plane detection to place objects at correct scale relative to the room.
5. Rendering. The 3D model is composited onto the camera feed using physically based shading, environment lighting estimation, and occlusion masks that hide parts of the product behind the user’s nose, hair, or hand. Frame rates target 30 to 60 fps to keep motion stable.
6. Interaction and capture. Shoppers swap colors, sizes, or styles, save snapshots, and share to social channels. Variant changes are mapped to product SKUs so that selections carry through to the cart.
7. Analytics and handoff. Events such as try-on start, variant switch, photo capture, and add-to-cart are logged to the retailer’s analytics stack. These signals support attribution, A/B tests, and merchandising decisions.

Common AR try-on categories

AR try-on is shaped by the body part or environment involved. Each category has its own tracking, sizing, and content needs.

• Eyewear. Sunglasses and prescription frames sit on face landmarks and rely on accurate temple width, lens curvature, and nose bridge geometry. Warby Parker offers a virtual try-on in its iOS app using the TrueDepth camera, and Lenskart, EssilorLuxottica brands, and Zenni operate similar experiences across web and mobile.
• Footwear. Sneakers and other shoes are anchored to feet using specialized foot trackers from vendors such as Wanna, Snap’s AR try-on for shoes, and Vyking. Brands including Gucci, Adidas, New Balance, and Puma have shipped foot try-on through their own apps and Snapchat Lenses.
• Furniture and home goods. Sofas, tables, rugs, and decor are placed on detected floor planes at one-to-one scale. IKEA Place, Wayfair’s View in Room, Houzz, Anthropologie, and Crate & Barrel use ARKit and ARCore so that shoppers can see whether an item fits the space.
• Beauty and cosmetics. Lipstick, eyeshadow, foundation, and nail polish are rendered on face meshes with shade matching and finish controls. L’Oréal’s ModiFace powers experiences for Maybelline, Lancôme, and Garnier, while MAC, Sephora’s Virtual Artist, NYX, and Charlotte Tilbury operate comparable tools.
• Watches and jewelry. Watches sit on the wrist using hand tracking or printed reference markers; rings use finger tracking; earrings and necklaces use face and neck landmarks. Cartier, Rolex retailers, Pandora, and Kendra Scott have launched experiences in this category.
• Apparel and accessories. Hats, scarves, handbags, and selected garments are previewed using body, head, or hand tracking. Full-body apparel remains a harder problem because of cloth simulation and pose variability, and many retailers use hybrid approaches that combine AR with 2D image overlays or photo-based virtual try-on.
• Automotive and large goods. Vehicles, appliances, and outdoor equipment are placed at scale in driveways or showrooms. Porsche, BMW, and Hyundai have shipped AR configurators that let prospective buyers walk around a car parked virtually in front of them.

Across these categories, published case studies report that AR try-on lifts conversion rate and reduces return rate. Shopify has reported that products with AR or 3D content can convert at materially higher rates than products without, with figures often cited around a 94 percent uplift in conversion in merchant case studies. Vendors including Wanna, ModiFace, and Threekit have published return-rate reductions in the range of 20 to 40 percent for participating retailers. These numbers vary by category, traffic source, and baseline, so retailers tend to validate them through their own holdout tests.

See also

Augmented reality (AR) — the broader category of technology that overlays digital content onto a live view of the physical world, of which try-on is one commercial application.

3D product configurator — an interactive tool for customizing a product in 3D, often paired with AR try-on so that a configured variant can be previewed in context.

WebXR — the browser API that exposes AR and VR capabilities to web pages, used to deliver AR try-on without requiring an app install.

Spatial computing — the wider field of computing that blends the digital and physical worlds, providing the platforms and devices on which AR try-on runs.

Additional resources

• Apple ARKit — documentation for the iOS framework that powers face, world, and body tracking used in many AR try-on apps.
• Google ARCore — documentation for the cross-platform framework that supports plane detection, augmented faces, and depth on Android and web.
• Google model-viewer — an open source web component for rendering interactive 3D models with built-in AR quick-look support.
• WebXR Device API specification — the W3C draft that defines how browsers expose AR and VR sessions to web applications.
• Snap AR — Lens Studio, Camera Kit, and AR try-on tooling used by retailers to ship experiences inside Snapchat and partner apps.
• Shopify on AR shopping — merchant-focused write-ups on AR and 3D commerce, including conversion case studies for AR-enabled products.