I came to Skinive AI for the CE mark. Most consumer skin scanners run on internal validation and confident percentages. Skinive cleared a real regulatory bar in the EU as a Class I medical device under MDR, which is the kind of detail most reviewers ignore and most readers should pay attention to. Sixteen days in, with 19 mole scans, three suspect spots, and a week of acne tracking layered on top, I have a more nuanced view than the CE badge alone suggests.
TL;DR
Regulated, broad in scope, sharper on darker skin tones than most competitors, and conservative in the right direction on melanoma-adjacent reads. The dermAI score is useful as a flag and easy to misread as a verdict. The 5M analyses claim is impressive and also not the right way to think about model quality.
What Skinive AI is
Skinive AI is a mobile app from a Netherlands-based team, available in a consumer build and a professional Skinive MD variant. The consumer app scans photos against more than 50 skin conditions, including melanoma-adjacent pigmented lesions, basal cell carcinoma indicators, acne, rosacea, eczema, fungal infections, and inflammatory dermatoses. Each scan returns a dermAI risk score, a top differential, and a recommendation tier from low concern to immediate clinical follow-up. The CE mark places it inside the EU MDR framework as a Class I device with a defined intended use. The company cites more than five million analyses performed and more than 130,000 oncology-risk cases identified across its consumer and clinical channels.
Who it’s for
Readers in the EU who want a regulated, documented triage tool between annual derm visits. Anyone with a personal or family history of skin cancer who needs a structured way to monitor known lesions month to month. Slow-skincare readers with chronic inflammatory skin conditions who want a tracking layer that recognizes their condition family rather than scoring everything as an undefined hazard. Clinicians or aestheticians who want a documented AI second-opinion alongside their own exam, the Skinive MD variant is built around that workflow.
Skip it if you want product and routine analysis, that is a different category and Skinive does not pretend to do it. Skip it if you are in a region where the CE mark does not map to your local regulatory framework, and you need a locally-cleared tool. Skip it if you cannot tolerate a screening tool that asks you to follow up rather than giving you a clean yes or no, the conservatism is the feature.
Features that matter
- CE-Marked Class I medical device. Under EU MDR, this means documented intended use, post-market surveillance, and clinical evaluation. Most consumer scanners do not clear this bar.
- 50-plus condition coverage. Broader than mole-only apps, more conservative than scanners that claim to recognize hundreds of conditions on thin data.
- dermAI risk score. The output that flags concern level and recommended action. Useful as a flag, not as a diagnosis.
- Cross skin-type validation. The training data and validation cover a wider range of Fitzpatrick types than most competitors. This is the feature I would lead with if I were rewriting their App Store copy.
- Skinive MD professional variant. A separate workflow for clinicians, which gives the consumer model the benefit of clinical feedback loops most consumer scanners do not have access to.
My contrarian take
The 5 million analyses claim does a lot of marketing work and not enough scientific work. A scanner can run 50 million analyses and still have poor recall on the lesions that actually matter. The headline figure tells you nothing about the false-negative rate on melanoma, which is the number that should drive the install decision. The same goes for the 130,000 oncology-risk cases identified, which is impressive in volume and silent on follow-up rates and confirmed diagnoses. The CE mark is the real evidence here, because it forces documented intended use and clinical evaluation. Read the published validation work and look for the per-condition breakdown, especially for pigmented lesions, before assigning the dermAI score more weight than it earned.
Real-world test
I tested Skinive AI for 16 days in late April and early May, with the test bracket centered on a planned dermatology visit on day 14 so I could compare the app’s reads against a clinician’s assessment of the same lesions. I scanned 19 known moles across my back, shoulders, and right thigh, with my partner taking the back photos under a north-facing window between 10 and 11 in the morning. I added three suspect spots that had appeared over the previous six months.
The 19 known moles all returned low-concern dermAI scores. Two of the three suspect spots returned medium-concern, with a recommendation to follow up within 30 days. The third returned low with a note about reassessing in three months. On day 14, the dermatologist confirmed the low reads on the 19 known moles, agreed with the medium reads on both suspect spots and biopsied one, and downgraded the third to no concern. The biopsy came back benign, which I would not have predicted from the app alone, and the dermatologist said the biopsy was the right call given the visual presentation. That is the bar for a triage app working as intended.
How it compares
ScanSkinAI is the closest direct competitor on accuracy claims, with an independent validation paper that ScanSkinAI cites and Skinive does not match on that exact axis. Skinive’s CE mark is a stricter regulatory bar overall. First Derm has the strongest teledermatology bridge if you want a real clinician in the loop within 24 hours, which Skinive does not offer in the consumer build. For pure product and cosmetic analysis, Skinive is not the tool, look at Lovi or Cosmily. If your reason to install is high-stakes triage with regulatory cover, Skinive is the strongest pick in this group for EU readers. For US readers, the comparison is more even and the answer depends on whether the CE mark maps to your local risk tolerance.
FAQs
What does CE-Marked Class I actually mean? It means the app is registered under EU MDR as a Class I medical device with documented intended use and post-market surveillance obligations. It is not a guarantee of accuracy on every scan. It is a regulatory bar that most consumer scanners do not clear.
Should I rely on the dermAI score? Treat it as a flag, not a verdict. Low-concern is a reassurance, not a clearance. Medium and high are prompts to see a clinician.
How does it perform across skin tones? Better than most consumer scanners, on the validation data the company publishes. Representation is still uneven across all Fitzpatrick types, which is true of every consumer scanner.
What about the professional Skinive MD version? It is a separate workflow for clinicians and aestheticians. Most readers will not need it. Knowing it exists explains some of the consumer app’s design choices.
Is my data shared? The CE mark forces a documented data-handling posture. Read the current privacy policy before installing, and look specifically at whether anonymized scans are used to retrain the model and whether you can opt out.