Notes

Makeup try-on apps compared: what each one shows you

Last updated: 22 August 2026

Makeup try-on splits into five kinds of tool, and each one shows you something the others cannot.

Makeup try-on splits into five kinds of tool, and each one shows you something the others cannot. A brand's AR filter tracks your face live. A shade database only knows what someone else already bought. Generative try-on, the category Looksmith sits in, hands you a picture of a whole look on your own face, about a minute after you ask for it.

This is a map of the five categories, not a ranking. If a brand's own AR filter already answers your question, that is the right tool, and nothing below argues otherwise.

Five kinds of tool, not one industry

The apps and devices people lump together as "makeup try-on" run on different hardware, come from different companies, and were built to answer different questions. Sorting the actual products into their category first stops you from comparing two things that were never solving the same problem.

Brand-owned AR, on the brand's own product page. Most of what you have used on a brand's own site is a licensed engine wearing that brand's name. L'Oréal acquired ModiFace, a company that builds this kind of augmented-reality beauty technology, on 16 March 2018. The result is a page that can move its own shades with your face and nothing it does not sell.

Standalone AR apps you install yourself. Same live tracking, wider catalogue, not tied to one brand's shelf. Perfect Corp, which makes YouCam Makeup, has been a publicly traded company on the New York Stock Exchange since 2022. Meitu's BeautyCam sits in the same category, and its own App Store listing puts "Trendy Makeup" beside "AI Slimming" and "3D Nose Sculpting" in the same feature list, features that have nothing to do with a lipstick shade.

In-store devices, operated for you. Sephora's Color IQ is a handheld scanner an associate runs across your skin, and it produces no image at all. According to the original Sephora and Pantone announcement, it launched on 26 July 2012. It assigns a Pantone SkinTone number from measured reflected light and returns a list of foundations that match it.

Swatch photography and shade databases. Sites such as Findation carry no camera and no scanner. Findation describes itself as a database built from other people's purchases: you submit a shade that worked for you, and it suggests the same shade in a different brand, drawing on a database it puts at over 11,000 products and 106,000 shades. Nothing about the recommendation involves your face. It involves a stranger's, and how closely that stranger's skin, camera and screen happen to match yours.

Generative try-on. Instead of tracking or scanning, a model regenerates the photograph itself, and the category is not one product. Doji's own App Store listing describes trying on fashion items from an AI likeness, not makeup. Looksmith does the same thing for makeup: you save a Likeness once, and Looksmith paints a complete look, lips through brows, onto your own photograph rather than someone else's.

Category by category

Which of these measures your skin instead of guessing?

Only one does: the in-store scanner. Color IQ and devices like it read reflected light off your skin and assign a number to it. Every other category here, including generative try-on, is either painting colour onto a live image or predicting how a colour would look. A measurement beats a guess when the question is a single shade. It says nothing about a whole look, because a scanner was never built to answer that.

Which one can I keep, compare or send to someone?

Only the categories with a physical output do that reliably: the product list a store associate hands you, and the image file generative try-on returns. Both kinds of AR, brand-owned and standalone, are a live overlay. Close the app or the browser tab and it is gone unless you thought to screenshot it, and a screenshot of a moving overlay is a worse copy of a thing that generative try-on was built to produce in the first place. Looksmith returns a 9:16 poster and, optionally, a four-second clip, with the products used listed beside it.

Does any of these show a whole look instead of one product at a time?

Generative try-on is the only category built for that. Brand-owned AR is scoped to one brand's catalogue by definition. Standalone AR apps can cycle through several products in one session, but you are still watching a live video try each one, not looking at a finished, composed face. A shade database and an in-store scanner both answer a single-product question by design; asking either one about a whole look is asking it something it was never built to do. Looksmith builds a complete look across twelve regions, lips through brows, at once.

Will any of these change something on my face I did not ask about?

Check the app rather than assume. A brand's own AR page is usually scoped narrowly enough that it will not, because reshaping is not what sells that brand's lipstick. Standalone apps are a different case: BeautyCam's own listing puts "AI Slimming" and "3D Nose Sculpting" in the same feature list as its makeup looks, and a setting left on in one session is easy to miss in the next. Looksmith does not do this by design. It will not slim, lift, resize or reshape anything; a monolid stays a monolid, and the pose, age and expression in the photograph you sent are the pose, age and expression that come back.

Where does Looksmith fit in this list?

In the last row, as the only generative try-on app in this comparison built specifically for makeup, coming soon to iOS. In Looksmith you save a Likeness once, a set of selfies that becomes your reusable face, then reach a look one of three ways: Style Me puts you in several looks at once, Create Look takes a description in words, and Mirror a Look transfers the makeup from a photograph you found somewhere else. The first six looks in Looksmith are free, granted at onboarding; after that Looksmith uses credits, bought in packs, and there is no subscription. What happens to the photograph itself is in the privacy policy.

The question that decides it

Not which is best, because the five categories on this page are not competing for the same job. Ask what you are trying to find out. A single shade, in a hurry: use the app already on the brand's page, or the in-store scanner if the shop has one, because a measurement or a live match beats waiting a minute for a picture. A whole look, and something worth looking at again tomorrow: that is what generative try-on is for, because it is the only category that hands you a finished image instead of a live overlay or a stranger's old purchase. If the AR filter already answers your question, use the AR filter. That is not a hedge. It is the actual answer for a large share of the people who land on this page, and saying otherwise would just be marketing with extra steps.