makeup look curation for virtual try-on
Makeup Look Curation for Virtual Try-On: Building a Complete Preview With 13 Products
A complete preview brings lipstick, foundation, blush, eyes and contour together in one render. See where it pays off, what belongs to the SDK and what belongs to the API, and how to build looks that reach the cart.
Trying on a lipstick settles a doubt. Building a whole face settles a purchase. That is the shift makeup look curation for virtual try-on brings to makeup brands: instead of testing one isolated SKU, the shopper sees the full combination — foundation, blush, eyes, lips, contour — and decides on a set of products at once. For e-commerce and CRM teams, that changes what you put on screen and what you measure afterwards.
What makeup look curation for virtual try-on looks like in practice
MaIA Try-On comes in two forms, and they do not do the same thing.
The API is photo-based: the brand sends an image and a product — lipstick or foundation — and gets the rendered image back, with palettes calibrated to the real SKUs in the project. It is the path for server-side flows, WhatsApp, email, product pages, any channel where the rendering has to happen outside the browser.
The 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 the Full Look, which combines those items into a single preview. It runs in live camera mode, with rendering done locally in the browser, or in photo mode, accepts any #RRGGBB colour and is self-hosted by the brand.
In other words: the Full Look is an SDK feature. The API remains the photo-based path for lipstick and foundation. If you are deciding between the two, it is worth reading the comparison between photo-based try-on and live camera mode before you design the integration.
Where a complete look is worth more than a single product
Collection launch
Collections are sold as a concept: an autumn palette, a summer line, a collab. The Full Look shows the concept on the shopper's face instead of on the campaign model. The question stops being "does this lipstick suit me?" and becomes "does this look suit me?".
Kits and bundles
When the goal is to raise items per order, the look is the best kit display there is. Each product in the look becomes an addable SKU, and the cart button can carry the whole set, with the option to remove whatever the shopper already owns.
Content and campaign
A named look — "daytime glow", "evening with a strong eye" — works as reusable content in newsletters, social and category pages. The same look can be recreated with different SKUs for different skin tone audiences.
Guided replenishment
For the installed base, the look works as a visual reminder of what is missing: the shopper bought foundation and lipstick, never bought blush, and sees in the preview the effect of the item she does not have.
How to build a Full Look that works
Start with curation, not with technology. A look with thirteen products active at once is usually worse than a look with five well-chosen ones. Define five to seven items per look, with a clear focal point (lips or eyes), and leave the rest in a supporting role.
Treat the palette as catalog, not as a generic swatch set. The SDK accepts any hexadecimal colour, which is freedom and trap at the same time. If the hex used in the preview is not the calibrated shade of your SKU, the shopper sees one colour and receives another. Keep a single source of truth between the catalog and the values sent to the render, and version it when the formula changes.
Order the layers the way a makeup artist would. Foundation and concealer before blush and contour; highlighter after powder; lip liner before lipstick. Order affects the visual result and the perception of realism.
Keep the look editable. The value of the Full Look is not only the preset: it is the shopper swapping the lipstick inside the look and watching the whole set recompose. Allow one item to be changed without tearing down the entire preview.
Adapt the look to the profile. When the preview starts from a beauty profile — skin tone and undertone, skin type — the presets stop being generic. The selfie-based skin analysis API returns those attributes as JSON and lets the same look ship in different versions by undertone. The analysis describes visible attributes of the photo; it is not a medical assessment and does not replace a professional.
What to measure
A complete look calls for its own metrics, different from those of a single-product try-on:
- Full Look sessions started and completed;
- average number of items swapped inside the look;
- items added to cart per look session;
- which product in the look is removed most often (a sign of wrong curation or of a shade range off for that audience);
- which looks generate the most additions per category.
This data also feeds merchandising: if a dark-lip look is started a lot and converts little, the problem may be in the available shade range, not in the interface. Avoid comparing your results with numbers published by third parties — performance varies with each brand's catalog, traffic and channel.
Governance, privacy and timeline expectations
In the SDK's live mode, rendering happens locally in the browser and nothing is uploaded. In photo mode and in the API, the image goes through the flow the brand defined — and that needs to be described in the privacy policy and in the on-screen copy, not buried in a footer.
Two expectations to align internally before promising a date: the SDK is in early access, granted per account, with no public download; and the Full Look, with the thirteen products, is a browser feature — the photo-based API renders lipstick and foundation. Planning the roadmap with that split clear avoids rework. The MaIA Try-On page details both delivery forms.
Want to test a complete look with your own shades? Create an account, buy credits and generate your API key in the documentation — or talk to the team to assess access to the browser SDK.
Frequently asked questions
What is the Full Look in virtual try-on?
It is the combination of several makeup products in a single preview — foundation, blush, eyes, lips, contour and others — instead of testing one isolated SKU. In MaIA Try-On, the Full Look is a feature of the browser SDK, which covers thirteen products.
Does the MaIA Try-On API also render the Full Look?
No. The API is photo-based and renders lipstick and foundation, with palettes calibrated to the project's SKUs. The other eleven products and the Full Look are features of the browser SDK.
How many products should I put in a look?
In practice, five to seven well-chosen items work better than switching everything on at once. Set a focal point (lips or eyes) and use the rest as support, letting the shopper swap items inside the look.
Are customers' photos uploaded to a server?
In the SDK's live mode, rendering happens locally in the browser and nothing is uploaded. In photo mode and in the API, the image follows the flow the brand defined, which should be described in the privacy policy and on the screen itself.
How do I get access to the browser SDK?
The SDK has been in early access since September 2026 and is granted per account, with no public download. For photo-based lipstick and foundation try-on, you just create an account, buy credits and generate the API key.