Does in-store virtual try-on replace physical testers?

No — it changes the order. The screen narrows a wide shade range down to one or two candidates in seconds, and the physical test is reserved for the final choice, which saves time at the counter and avoids applying and removing several products.

Do customer photos leave the tablet?

In the MaIA Try-On browser SDK's Live mode, rendering happens locally in the browser and nothing is uploaded. The photo-based MaIA Try-On API works differently: you send an image and receive the rendered image back, which is the right choice when the customer wants a picture to take away.

How do we make the on-screen shades match the physical product?

Through calibration. In the API, palettes are configured per project against the brand's real products. In the SDK you pass any #RRGGBB value, so the hex codes should come from the same source of truth as your packaging and product pages, with identical shade names across screen and shelf.

Can we use skin analysis to pick the starting shades?

Yes. MaIA Scan returns a beauty profile from a selfie — skin tone and undertone, skin type and condition, hair attributes, with a confidence score per field — which you can use to open the tablet on a shortlist instead of the full range. It describes visible attributes of a photo and does not diagnose anything.

In-Store Virtual Try-On at the Beauty Counter

Q: Which makeup products can be previewed in store?

The browser SDK covers thirteen products — lipstick, foundation, blush, face powder, under-eye concealer, eyeshadow, mascara, eyebrow makeup, eyeliner, bronzer, contour, highlighter and lip liner — plus a combined Full Look. The API renders lipstick and foundation from a photo.

The physical counter is still where a lot of makeup gets sold, and it is also where the oldest friction in beauty lives: the tester. Shared applicators, limited shade displays, a queue of customers waiting for the one consultant who knows the foundation range by heart. In-store virtual try-on moves that first step to a screen — the customer sees a lipstick or a foundation on their own face in seconds, the consultant keeps their hands free, and the shade conversation starts from something visual instead of from memory.

This guide is about the operational side: what to put on the tablet, how the staff workflow changes, and which decisions you need to settle before the first store goes live.

Why in-store virtual try-on behaves differently from e-commerce

Online, try-on competes with doubt. The shopper cannot touch anything, so a preview is often the only sensory input available. In store, the product is right there — so the preview has to earn its place for other reasons.

Three reasons that actually hold up in a store

Shade breadth. A counter displays what fits on the counter. The catalog is always larger. A screen can show every shade in the range, including the ones that never made it into the physical display, and including shades that are out of stock in that store but available for delivery.

Speed. Removing and reapplying a lipstick between three shades takes minutes. Tapping between three shades takes seconds. The consultant can narrow the range on screen and reserve the physical test for the final one or two candidates.

Hygiene and comfort. Not every customer wants to apply a shared tester, and not everyone is comfortable being made up in public. A preview is a lower-commitment entry point into the conversation.

What you put on the tablet

There are two technical shapes for in-store virtual try-on, and they fit different store formats.

Live camera mode on the brand's own device

The MaIA Try-On browser SDK, in early access since September 2026, covers thirteen products — lipstick, foundation, blush, face powder, under-eye concealer, eyeshadow, mascara, eyebrow makeup, eyeliner, bronzer, contour, highlighter and lip liner — plus a combined Full Look. It runs in Live mode, using the device camera with rendering done locally in the browser, so nothing is uploaded, or in Photo mode, from a single picture. It accepts any #RRGGBB colour and is self-hosted by the brand, which means the in-store page is yours: your shade names, your layout, your offline-tolerant hosting setup.

Live mode is the natural fit for a counter: the customer sees themselves move, the consultant can point at the screen, and because the rendering stays on the device, the privacy conversation at the store is simple — no image leaves the tablet.

Photo-based try-on through the API

The MaIA Try-On API works differently: you send a photo plus the product, and the rendered image comes back. It covers lipstick and foundation, with palettes calibrated to each brand's real products in a project. That is the right tool when you want a shareable output — an image the customer can receive on WhatsApp, keep, and look at again at home, or that the consultant can attach to a CRM record. It is photo-based, not live.

Many in-store setups end up using both: live mode for exploring at the counter, a photo-based render for the one or two shades the customer wants to take away. If you are still weighing the two approaches, we compared them in detail in photo-based virtual try-on vs AR try-on.

Making the shades match what is on the shelf

A preview that shows a shade slightly off from the physical product damages trust faster than having no preview at all — and in store, the customer can check immediately.

Two practical rules:

Calibrate against the real products, not against the marketing image. In the API, palettes are configured per project against the brand's actual products. For the SDK, the hex values you pass are your responsibility: pull them from the same source of truth your packaging and PDPs use.
Keep shade names identical everywhere. If the screen says "Nude 04" and the box says "Rosé Nude", the consultant spends the next minute translating instead of selling. Shade naming is a catalog hygiene problem, not a technology problem, and it is worth fixing before the pilot.

Adding a beauty profile to the counter conversation

Try-on answers "how does this look on me". It does not answer "where should I start". That is where selfie-based analysis comes in: the beauty scan API takes a photo and returns a beauty profile as JSON — skin tone and undertone, skin type and condition, hair attributes, with a confidence score per field. It is trained on Brazilian skin, which matters in a market where shade ranges have historically been calibrated elsewhere.

In a store flow, that profile narrows the starting set: instead of scrolling forty foundation shades, the tablet opens on the handful consistent with the detected tone and undertone, and the customer tries from there. Note what this is and is not — it describes visible attributes of a photo. It does not diagnose skin conditions and it does not replace a professional.

If you want the narrowing to be catalog-aware and conversational, MaIA Beauty Advisor is delivered as a project with the B4A team and reads the profile to recommend your own products across WhatsApp, product pages and chat.

Before the first store: five decisions

Who holds the device — customer-operated kiosk or consultant-operated tool? They need different screen layouts.
Connectivity — store Wi-Fi is unreliable. Decide what the page does when the network drops.
What happens after the preview — nothing, a WhatsApp send, a CRM record, a basket? Decide before launch, not after.
Staff training — five minutes on the tablet, and a script for the handoff from screen to physical test.
Measurement — pick one or two stores as a pilot and compare against similar stores without the tablet. Any conversion figures you see published belong to the brand that published them; your results will depend on your catalog, your traffic and your store format.

Start small. One store format, one product category, one clear next step after the preview.

Ready to build it? Create an account and get an API key to start testing, or talk to the B4A team about SDK early access for your store pilot.

In-Store Virtual Try-On: Makeup Previews at the Beauty Counter Without a Tester