Virtual Makeup Try-On for Lipstick: How to Present the Whole Shade Range Without a Tester

Lipstick is, almost always, the category with the most colour SKUs in a beauty brand's catalog — and the one that suffers most from the gap between the product photo on the site and the result on each person's lips. A well-built virtual makeup try-on for lipstick solves exactly that point: instead of asking the shopper to imagine how that red behaves on her skin, you show her. This article covers how to build that experience in a practical way, with the brand's real shade range, and where it makes a difference in the funnel.

Why lipstick is a different case from foundation

With foundation, the doubt is binary and painful: get the shade right or wrong. With lipstick, the doubt is about repertoire. The shopper rarely has a single right shade — she has dozens of possible shades and needs to decide which one to buy now. The obstacle is not the fear of a glaring mistake, it is paralysis in front of a grid of forty little coloured squares that say nothing about how they look on brown skin, on very fair skin or with an olive undertone.

That changes the design of the experience. In a lipstick virtual try-on, the value lies in allowing quick comparison between several shades, not in a single perfect simulation. The internal metric that matters is how many shades the person tries before choosing, and whether she reaches a decision instead of abandoning the page.

The problem with the physical tester

In retail, the lipstick tester has always been a sensitive point for hygiene and replenishment. In digital, it simply does not exist: what is left are swatches on a model's arm, which represent neither the colour on the lips nor on the shopper's skin. Virtual try-on fills that gap with no marginal cost per shade tested.

Two ways to deliver the virtual makeup try-on for lipstick

MaIA Try-On comes in two forms, and they serve different scenarios.

Photo API

The brand sends a photo and the product — lipstick or foundation — is rendered onto it; the image comes back ready. The shade ranges are calibrated to the brand's real products within a project, meaning the shade that appears in the simulation matches the SKU in stock, not a generic colour.

This format is the simplest to fit into channels where a photo is already circulating: a WhatsApp conversation, a CRM flow, an email with a personalised image, a human service agent who receives the shopper's selfie. Because rendering happens on the server, there is no requirement for a camera, permissions or device performance — and the result is an image, easy to save, resend and share.

Browser SDK, in early access

The SDK runs in the browser and covers thirteen products, among them lipstick, lip liner and a combined Full Look. It works in live mode through the camera — rendering happens locally, nothing is sent — or in photo mode, with any colour in #RRGGBB, hosted by the brand itself. It is the right format for a product page where the shopper will switch between many shades in a few seconds.

It is worth keeping the two clearly separate when planning: Full Look, lip liner and the other eleven products are exclusive to the SDK and work as a local preview. The photo API covers lipstick and foundation. If you want to understand better when each approach pays off, we wrote a direct comparison in photo try-on vs. AR try-on. The SDK has been in early access since September 2026 and is granted per account — there is no public download.

How to narrow the shade range before showing it

Showing forty shades is better than showing none, but showing eight relevant shades is better than forty. This is where virtual try-on stops being a toy and becomes a sales tool.

Selfie skin analysis returns, in JSON, visible attributes such as skin tone, undertone, skin type and condition and hair characteristics, with a confidence level per field. With the undertone in hand, you can order the lipstick range before the shopper touches anything: open with the shades that work with that undertone, leave the rest in a second block. The person can still test everything, but she starts from a point that makes sense.

The usual caveat applies: this describes visible attributes of a photo, for cosmetic recommendation purposes. It is not a clinical assessment and does not replace a health professional.

From test to recommendation

Those who want to close the loop — profile, recommendation and conversation — usually connect the virtual try-on to an advisor integrated with the catalog, which reads the beauty profile and suggests the brand's own products. That is a project run with the B4A team, not a self-serve API.

What to measure after launch

Before requesting a quote, define what you will observe. A few questions that tend to separate a useful pilot from a pretty showcase:

• How many shades are tested per session, and in which channels.
• Which shades are tested most and never bought — a sign of frustrated expectations or of price.
• What the difference in add-to-cart rate is between those who tested and those who did not, in the same period and the same traffic.
• Whether the volume of colour-related exchanges changes over a few purchase cycles.

Results vary with each brand's catalog, traffic and channel. Start with one lipstick line, not the whole catalog: calibrating twenty shades well teaches more than calibrating two hundred badly.

Start small and calibrated

A lipstick virtual try-on delivers value when the colour on screen is the colour of the SKU and when the shopper reaches it with the range already narrowed. The rest is integration.

Create your account, generate the API key and test the rendering with your own shades in the API documentation — or talk to the team to discuss a project with a calibrated shade range and SDK access.