Haut.AI Skin Analysis: How Brand Skin Scanners Work

Haut.AI skin analysis is a white-label engine that beauty brands embed in their own websites, apps and in-store tablets. You take a selfie, and the software returns scores for signs such as lines, pores, redness, pigmentation and hydration, plus a "perceived age." It powers Neutrogena's refreshed Skin360 tool. The accuracy figures you will see come mostly from Haut.AI itself.

Below we explain what the engine measures, where you are likely to meet it under another brand's name, what the published evidence does and doesn't show, and how to get a result worth looking at. This is a review of public documentation and research, not a hands-on test.

Is there a better alternative?

It depends on what you want from the scan. A brand scanner is built to end in that brand's product recommendations. If you would rather track your own skin over time and check any product's ingredients against your skin type, consider Pori. We make it, so weigh that accordingly. It is an iPhone app (iOS 17 or later) that turns a selfie into a 0 to 100 skin score with five aspects (Hydration & Moisture, Pores & Texture, Pigmentation & Tone, Anti-aging & Elasticity, Overall health) and a skin-age estimate in about 10 seconds. Its ingredient scanner (Pro) flags each ingredient as beneficial, use with caution or avoid for your skin type using fixed rules over about 250 INCI ingredients. Scan history stays on your phone. The first scan is free; Pro is $9.99 a week with a 3-day trial for eligible Apple IDs, or $49.99 a year (US, as of October 2026).

Haut.AI still has real advantages: it works in a browser with no download, it usually costs shoppers nothing, and its camera checks lighting and head position before it takes the photo. The rest of this review looks at it on its own terms.

What is Haut.AI?

Haut.AI is a business-to-business software company. You don't download a "Haut.AI app." Brands license the technology and put their own name on it. Its AI Skin Analysis product page describes an embeddable skin consultant app that a brand adds to its site, links that can be shared by QR code, email or in-store tablet, and a server-to-server API for apps and e-commerce platforms. Brands set their own branding, product catalog and recommendation logic in a management portal.

The same company sells a wider set of tools listed on its homepage: an AI skincare chat, SkinGPT (which simulates how products might change the look of skin), hair and body analysis, and software for clinical studies. The consumer-facing piece most people encounter is the selfie skin analysis.

What does a Haut.AI skin analysis measure?

The current engine, called Face Skin Analysis 3.0, returns about 16 outputs. Haut.AI's developer documentation lists them and how each is scored.

A brand rarely shows you all of these. It picks a subset, renames them and often rolls them into its own overall score. Neutrogena's version, for example, shows eight indicators and a single score out of 10 (more on that below). So if two brand sites both run on Haut.AI, their results pages can still look quite different.

The marketing numbers vary too. The product page talks about "150+ multidimensional facial biomarkers," while the homepage says "20+ skin health and beauty metrics." Those are different ways of counting the same system, so don't read much into the headline figure.

Where will you meet Haut.AI?

Usually without seeing the name. The clearest public example is Neutrogena. In September 2025 Haut.AI announced it had refreshed Neutrogena Skin360. The updated tool reads eight indicators (hydration, smoothness, tone evenness, radiance, firmness, dark spots, wrinkles and clarity), combines them into an overall score from 1 to 10, needs one live or uploaded selfie, and runs on Neutrogena's websites and selected retailers in the US and Canada. Our Neutrogena Skin360 review looks at that tool in detail.

Other relationships are public but less specific about what shoppers see. Ulta Beauty and Haut.AI announced a partnership in March 2022 to build skin analysis and recommendation technology together. The Haut.AI homepage also shows logos including Beiersdorf, NAOS Bioderma, Noom and Grupo Boticário. A logo tells you there is a business relationship; it doesn't tell you which consumer tool, if any, uses the engine.

A practical tip: if a brand's scanner guides you with on-screen prompts about head position and lighting before it takes the photo, it may be using Haut.AI's capture component, called LIQA.

How accurate is Haut.AI skin analysis?

Nobody outside the company has published a validation of the current consumer engine that we could find, so its accuracy hasn't been independently proven. Here is what exists.

• Vendor claims. The product page claims "98% clinical-grade diagnostic accuracy" against expert-labeled data and says the models were trained on over 3 million images. Neutrogena coverage in Personal Care Insights puts it at over 3 million images from around 800,000 people. The underlying study for the 98% figure isn't published on the page.
• Clinical software claims. For its research product, Haut.AI reports ICC values of 0.97 to 0.98, a measure of repeatability, across clinic cameras, clinic smartphones and at-home selfies. The partners and full methods aren't named on the page.
• A peer-reviewed age model. In 2018, Haut.AI researchers and partners published PhotoAgeClock, a model that predicted age from high-resolution photos of the eye corner with a mean error of 2.3 years. It was trained on 8,414 images of Caucasian women aged 20 to 80, and the authors note it works best on standardized photos. That is an older research model with a narrow population, not the selfie tool on a brand site.
• A study using its metrics. A 2022 paper in Plastic and Reconstructive Surgery used Haut.AI's skin metrics software on 433 selfies and found that most scores differed by sex and age, with wrinkles most strongly tied to chronological age. Its author sits on Haut.AI's management board, and it was about scoring scales, not accuracy.

None of this means the engine is bad. It means you should treat a single result as an estimate and lean on trends across repeat scans. For a wider look at the evidence across tools, see is AI skin analysis accurate.

How does Haut.AI estimate skin age?

Its "Perceived Age" output predicts how old the skin on your face looks, from 18 to 100. According to the Perceived Age documentation, it reads aging-related features such as wrinkles, sagging and pigmentation from one to three photos. The documentation gives no error figure for this output. A separate "eyes age" output estimates the age of the eye area alone.

Perceived age is a judgment of appearance. It moves with lighting, expression and makeup as well as with your skin, so a result five years above your real age after one photo under an overhead light tells you little. Our guide to what skin age means covers what drives the number and what you can change.

Why does the photo step matter so much?

Because the analysis can only be as good as the image, and Haut.AI says so itself. LIQA, its "live image quality assurance" camera component, checks face visibility, position in the frame, and lighting intensity and direction in real time, with on-screen guides until the photo passes. That's a genuine strength over apps that accept any snapshot.

Two caveats come from the documentation. First, brands can relax the lighting check: LIQA's Required Lighting setting has an "adaptive" option that eases the requirement when light is poor and a "none" option that turns lighting validation off. Second, the analysis docs recommend images of at least 2,500 by 2,500 pixels with the face filling 85% of the frame, a clean face with no makeup and a neutral expression. An uploaded photo below 256 by 256 pixels is accepted but gives results that are "not meaningful." If a site lets you upload an old, filtered or small photo, the score you get back may not reflect your skin.

What are the strengths and limits of brand skin scanners?

What happens to your selfie?

Haut.AI's product page describes a "Skin Atlas" step that removes identifying features such as the eyes, hair and eyebrows before cloud processing, and states that no facial images are retained in the cloud. The brand running the tool is a separate company with its own privacy policy, and it may store your results, email address and product choices. Read the brand's privacy notice before you upload, and use the "upload" option only with a photo you're comfortable sharing. Our guide to skin analysis app privacy lists the questions to check.

How do you get a useful result from a brand skin scanner?

1. Wash your face and skip makeup, tinted sunscreen and heavy moisturizer for the scan.
2. Face a window in daylight so light falls evenly on both cheeks. Avoid overhead bulbs that cast shadows under the eyes.
3. Pull hair off your forehead, remove glasses and keep a neutral expression.
4. Use the live camera with on-screen guidance instead of uploading an old photo.
5. Repeat the scan in the same spot and light a few weeks later, and compare the trend instead of a single number.
6. Treat product suggestions as one brand's catalog. Check the ingredient list against your own skin type before you buy.

Is Haut.AI skin analysis a medical diagnosis?

No. Brand scanners describe how skin looks, for cosmetic and shopping purposes. They don't check moles or diagnose acne, rosacea or other conditions. If a spot is new, changing, bleeding or looks different from your other moles, the American Academy of Dermatology's ABCDE warning signs are a better starting point, and a dermatologist is the right person to see.