Eyewear try-on
Glasses and sunglasses fitted to a live face mesh, with real scale calibration so frame width reads honestly against the user's face rather than floating at a flattering default.
All services
Virtual try-on · AR commerce
Virtual try-on, built around your catalogue.
Real-time try-on that runs in the browser — face-tracked, catalogue-driven, and integrated with the store you already have.
★★★★☆ 4.5 on Clutch · 50+ projects · 20+ clients · since 2021
What we build
Try-on that survives a real catalogue.
Jewellery & watches
Hand and wrist tracking for rings, bracelets and watches — the harder tracking problem, where occlusion and scale matter more than lighting.
Catalogue pipeline
The unglamorous half: a repeatable path from your product data to try-on-ready 3D assets, so adding an SKU is a catalogue action and not a 3D modelling project.
Store integration
Try-on wired into Shopify, WooCommerce, BigCommerce or a custom storefront — variant-aware, so switching colour in the configurator switches the product you add to cart.
In-store kiosks
The same build running on a mirror or tablet in a physical store, where lighting is controlled and the hardware is known.
Analytics
Instrumentation for the only questions that matter: try-on rate, time in session, and whether try-on sessions convert or return differently from the rest of your traffic.
Virtual try-on is a 3D pipeline problem wearing a marketing hat. The demo — one well-modelled frame on one well-lit face — is the easy part. What decides whether it ships is everything behind it: how 300 SKUs get modelled and kept in sync with the catalogue, how the thing performs on a mid-range Android phone, and how it degrades when face tracking loses the user.
What actually makes try-on hard
The three failure modes we see are consistent, and none of them are about the AR itself.
- Catalogue scale — A pilot with 12 hand-modelled products tells you nothing about 800. Without a repeatable asset pipeline and a naming contract with the PIM, every new product becomes a manual 3D job and the feature quietly dies.
- Device reality — Face tracking plus a physically-shaded 3D asset is a real compute budget. If it is only tested on flagship hardware, a large share of your traffic gets a slideshow or a black frame.
- Fit credibility — Shoppers forgive imperfect lighting. They do not forgive a frame that is visibly the wrong width for the face. Scale calibration is what separates a toy from a tool that reduces returns.
Browser or app?
For retail try-on, the browser almost always wins: no install stands between a product page and the try-on, which is the entire point. Native only earns the call when you need depth-sensor accuracy or want try-on inside an existing app you already ship.
| WebAR (browser) | Native app | |
|---|---|---|
| Friction | Tap a link or scan a QR code | App Store install first |
| Reach | Any modern phone or laptop | Only users who install |
| Tracking depth | Camera-based face mesh | Depth sensors where available |
| Best for | Product pages, campaigns, in-store kiosks | Existing app with an engaged base |
Proof
Browser 3D, shipped at enterprise scale.
National Bank of Kuwait
The browser-3D pipeline try-on runs on
We have not shipped an eyewear try-on — we will not pretend otherwise. What we have shipped, repeatedly, is the hard part underneath it: real-time 3D delivered in a browser at enterprise scale. NBK Virtugate is a full virtual environment running in Unity WebGL inside a memory ceiling, built for staff on ordinary corporate machines with no install. The constraints that project lives under — asset budgets, load time, graceful degradation on weak hardware — are the same constraints that decide whether a try-on feature works on your customers' phones.
See the work →How we work
From catalogue audit to live storefront.
Catalogue audit
A free scoping call where we look at your actual product data: how many SKUs, what variants, what 3D or photography already exists, and how it is kept up to date. This is what sizes the project.
Scope & proposal
A fixed-fee or milestone proposal with the device support matrix, the asset pipeline plan, and a stated success metric — try-on rate, conversion delta, or return-rate reduction.
Pilot
One category, a real subset of your catalogue, running on your product page and tested on the devices your analytics say your customers actually use. Judge it before committing to the full range.
Pipeline & rollout
The asset pipeline that turns catalogue additions into try-on assets, then the full range, then the store integration and analytics.
Launch & lifetime support
Deployment, team handover, and ongoing support as devices, browsers and your catalogue change. Lifetime technical support on everything we ship.
Stack · platforms & tools
The technology behind it
Trusted by enterprises, institutions & studios
- National Bank of Kuwait
- Meta
- Road Safety Authority Ireland
- Cubicle Ninjas
- Nova Slides
- Seat Unique
- Immersive Exposure
- Work Spatial
- Veem
FAQ · questions
What XR buyers ask us
What is virtual try-on software?
Virtual try-on software lets a shopper see a product on themselves through their camera before buying. For eyewear it tracks a live face mesh and fits the frame to it at real scale; for jewellery and watches it tracks hands and wrists. The useful versions run in the browser, so there is no app install between your product page and the try-on, and they are driven by your product catalogue rather than a fixed set of hand-built demos.
Does virtual try-on need an app, or can it run in a browser?
For retail, the browser is almost always the right answer. WebAR runs from a link or a QR code on any modern phone, which matters because an app install is a conversion cliff on a product page. Native apps earn the call in two cases: when you need depth-sensor accuracy beyond what a camera face mesh gives you, or when try-on belongs inside an app you already ship to an engaged customer base.
How many products can it handle?
That depends entirely on your asset pipeline, which is why we audit the catalogue before quoting. A pilot with a dozen hand-modelled products is straightforward; several hundred SKUs that change seasonally is a different project, because the modelling has to become repeatable and tied to your product data. We scope the pipeline first, since it is the thing that decides whether the feature survives its first catalogue update.
How much does virtual try-on development cost?
We scope every project as fixed-fee or milestone work after a catalogue audit. A single-category pilot on a subset of your range is the usual starting point and sits well below a full rollout; the cost drivers are SKU count, how much usable 3D or photography you already have, tracking complexity (faces are easier than hands), and the depth of store integration. You get a detailed proposal with a success metric before any work starts.
Will it work on older phones?
Some of them, and we will tell you which before you commit. Face tracking plus a shaded 3D asset is a real compute budget, so we set a device support matrix during scoping based on your own analytics rather than a generic assumption, then build to that floor with a graceful fallback — usually a high-quality static preview — for devices below it.
Can it integrate with our Shopify or WooCommerce store?
Yes. We integrate with Shopify, WooCommerce, BigCommerce and custom storefronts, and the important part is variant awareness: the try-on and the cart have to agree, so changing colour or size in the try-on changes the product the shopper actually adds. We also instrument it so you can see whether try-on sessions convert differently from the rest of your traffic.
Do we own what you build?
You do. All custom code, 3D assets and configuration transfer to you on final acceptance under a standard work-for-hire clause. Third-party libraries stay under their own licenses, which we disclose in the proposal, and we can work entirely under your paperwork and NDAs.
Explore more
Other service lines
Let’s scope your project.
Tell us about your world — we’ll scope it, design it, and deliver it end to end. Honest timeline, budget range, and the right platform.