There is a specific category of skincare app I keep auditing for Elelaf, and SkinGenie is the cleanest current example of it. The format is: take a selfie, answer quiz questions, get a personalized routine. The marketing copy talks about AI, computer vision, multi-factor skin analysis. The practical reality is closer to a decision tree with a photo upload bolted to the front. I am not saying that as a takedown. The routines that come out the other side are not bad. The difference between ‘algorithm built me a routine’ and ‘a flowchart that I could have written in a notebook built me a routine’ actually matters when you are deciding whether to trust the output.
What SkinGenie is
SkinGenie is a free skincare routine generator available on the web and Android. You upload or take a selfie, the app runs what it calls a 9-factor skin analysis (texture, hydration, tone, sensitivity markers, pore visibility, redness, oiliness, dark spots, fine lines), you answer a follow-up quiz about lifestyle, climate, and product preferences, and the app outputs a personalized AM and PM routine. You can toggle between a Korean-style multi-step routine and a Western-style minimalist one, save multiple routines, and switch between them as your skin changes. Product recommendations span price tiers. The whole thing is free with no paywall on the routine itself.
Who it’s for
Genuinely useful for beginners. If you have no routine, no idea where to start, and want a defensible starter stack that will not destroy your barrier, SkinGenie is a reasonable on-ramp. It is also useful for people who want to compare a K-beauty routine against a Western one for the same skin profile, because the toggle lets you see both side by side from the same inputs.
Not the right tool if you already have a working cabinet and want the app to respect what you own. There is no inventory layer; the routine assumes you are buying from scratch. Also not the right tool if you have a complicated dermatologic situation (active rosacea flares, prescription tretinoin, perioral dermatitis, melasma under treatment). The questionnaire does not have the resolution to handle those cases responsibly, and the routine it suggests will be too generic to help.
Features that matter
- Selfie + 9-factor scan. The scan returns scores for each factor. The scores feel directionally right (it flagged my T-zone oiliness and the redness around my nose correctly) but the resolution is coarse. Think ‘mild, moderate, high’ rather than precise measurement.
- Lifestyle quiz. 20-something questions covering climate, sensitivity history, fragrance tolerance, budget, and product preferences. This is doing more of the routing work than the selfie is.
- Korean and Western routine variants. Same inputs, two outputs. The Korean version is heavier on hydrating layers (toner, essence, serum, moisturizer). The Western version compresses to fewer steps. Useful as an educational compare-and-contrast.
- Routine library. You can save multiple routines and set one as active. Helpful for seasonal switching or for keeping a stripped-back travel routine alongside your full one.
- Free. No paywall on the routine, no subscription pitch, no kit funnel. The business model is presumably affiliate links on the product recommendations, which is the least invasive option in this category.
My contrarian take
The marketing leans hard on ‘AI’ and ‘computer vision’ as differentiators. The honest read is that most personalization in routine generators like this happens at the questionnaire layer, not the photo layer. The selfie analysis is doing some work, probably less than you assume. If you took the photo step out and just answered the quiz, my guess is the routine you got back would be 80 percent the same. A well-designed quiz is a perfectly reasonable way to route a beginner toward a reasonable starter routine. It is just not magic, and calling it AI does a small disservice to readers who think they are getting something more bespoke than a decision tree. The output is also conservative by design. SkinGenie will never push you toward anything interesting, which is correct for a free public tool and a limit you should know about.
Real-world test
I tested SkinGenie for 18 days starting in late April, across a Seoul-to-Lisbon flight and a climate shift that always wrecks my skin for a week. I ran the selfie scan three times across the period to see how stable the analysis was. The scores shifted slightly between scans (oiliness moved one notch, redness held steady), which is either real skin variation or scan noise. I could not tell which.
The Korean-style routine it generated for me was a six-step PM and four-step AM, anchored by a low-pH cleanser, a hydrating toner, a niacinamide serum, a ceramide moisturizer, and SPF in the morning. Reasonable. Nothing I would have argued with. The Western-style toggle compressed this to a cleanser, serum, moisturizer, SPF stack, which is also fine and which I have seen suggested by half the dermatologists on TikTok. I followed the Korean routine for 12 days. My skin tolerated it. I did not see dramatic improvement, but I also did not break out. The routine library let me save a stripped-back travel version for the days when I was sleeping on couches and could not be bothered with six steps. That feature is quietly the most useful one in the app.
How it compares
Against Cosmily, SkinGenie is a different category. Cosmily is an ingredient checker for products you are already considering. SkinGenie is a routine generator for people who do not know what to consider yet. They pair well: use SkinGenie to get a starter routine, then use Cosmily to vet the specific products. Against MDacne, SkinGenie has no commercial layer (no MDacne-branded kit, no $9 trial), which is structurally cleaner. Against Skin Logger or other free routine trackers, SkinGenie is more opinionated; it tells you what to do, not just lets you log what you did. If you want a routine handed to you, SkinGenie. If you want to log your own, you need a different tool.
FAQs
Is the AI scan actually doing anything? It is doing something, but probably less than the marketing implies. The questionnaire is doing more of the routing work. The selfie is a confirmation layer more than a primary input.
Is it really free? Yes. No subscription, no kit. The business model is most likely affiliate links on product recommendations.
Should I use the Korean or Western routine? Try both, see which one you actually maintain. Adherence beats step count every time. A six-step routine you skip is worse than a three-step routine you do.
Is the photo data stored? Check the current privacy policy before uploading. Selfie-based skincare apps have a mixed history on data retention, and the policy can change.
Will SkinGenie tell me to see a dermatologist if something is wrong? Not reliably. The questionnaire does not have the resolution to catch dermatologic conditions that need professional care. If you suspect rosacea, perioral dermatitis, or anything that is not responding to gentle care, see a derm, not an app.
If you want the broader landscape, the routine-builders hub has the rest of the apps in this category tested this round. SkinGenie is a fine starting point. Just calibrate your expectations to ‘thoughtful quiz’ rather than ‘AI dermatologist in your pocket’ and the value lands cleanly.