Genuine Korean skincare. Cash on delivery — open the parcel, then pay.

Skin Advisor

The research behind the Skin Advisor

Every question, every ingredient tip and every "see a doctor" warning comes from the guidelines, studies and datasets below. Here is everyone whose work we learned from, with their licence.

48 ingredients · 132 studies

Clinical guidelines and safety rules

  • American Academy of Dermatology — Reynolds R. V. et al., J Am Acad Dermatol 90(5), 2024

    Used for: Acne ingredients (benzoyl peroxide, adapalene, salicylic acid) and when to see a doctor

    Licence: Journal article (cited, summarised in our own words)

  • National Institute for Health and Care Excellence (NICE), UK

    Used for: When to send someone to a dermatologist (urgent / refer / consider)

    Licence: NICE guidance (cited)

  • Pigmentary Disorders Society, India — Indian Journal of Dermatology

    Used for: Melasma and dark-spot advice, written for South Asian skin

    Licence: Open access article (cited)

  • Indian dermatologists' Delphi consensus — open access in PubMed Central

    Used for: Sunscreen advice for hot, humid climates

    Licence: Open access article (cited)

  • Cochrane

    Used for: How well melasma and acne treatments work, across many trials

    Licence: Abstracts (cited)

  • European Commission

    Used for: Safe limits: retinol max 0.3% on the face, alpha-arbutin max 2%

    Licence: EU law

  • European Commission

    Used for: Ingredient safety and strengths

    Licence: EU public documents

  • Cosmetic Ingredient Review, USA

    Used for: Ingredient safety

    Licence: Cited

  • U.S. Food and Drug Administration

    Used for: Sunscreen filters

    Licence: Public domain

  • Mercury in skin-lightening creams, Bangladesh (2025)

    In use

    Environment and Social Development Organization (ESDO), Bangladesh

    Used for: The fairness-cream question and the mercury warning

    Licence: NGO report (cited)

Where the questions come from

  • Misery L. et al., University of Brest, France; cut-offs: Legeas C. et al., Acta Dermato-Venereologica 2021

    Used for: Ten symptoms that measure sensitive skin; above 13 means sensitive

    Licence: Published scale (cited)

  • Fitzpatrick T. B., Archives of Dermatology 1988

    Used for: The question on whether your skin burns or tans in the sun

    Licence: Published scale (cited)

Skin image datasets

  • DermaCon-IN

    Used for testing

    Madarkar S. S. et al., NeurIPS 2025 Datasets and Benchmarks — Harvard Dataverse

    Used for: 5,450 images of Indian skin: we check how well our photo read works on brown skin and when it says to see a doctor

    Licence: CC BY-NC-SA 4.0

  • Fitzpatrick 17k

    Being added

    Groh M. et al., CVPR Workshops 2021

    Used for: Checking it works equally well across skin tones

    Licence: CC BY-NC-SA 3.0

  • ACNE04

    Being added

    Wu X. et al., Nankai University — ICCV 2019

    Used for: Grading acne as mild, moderate or severe

    Licence: Free for academic use (credited)

  • FFHQ-Wrinkle

    Being added

    Moon J. et al., 2024 (on NVIDIA's FFHQ faces)

    Used for: Finding fine lines and wrinkles

    Licence: CC BY-NC-SA 4.0

  • Ward A. et al., Google Research and Stanford Medicine — JAMA Network Open 2024

    Used for: Spotting when to see a doctor in everyday phone photos

    Licence: SCIN Data Use License

Tools and ingredient databases

  • Google

    Used for: Reading the photo (with your permission); our own rules make the final call

    Licence: Google API terms

  • Google

    Used for: Checking the photo is clear and finding each part of the face

    Licence: Apache 2.0

  • Monk E., Harvard University (with Google)

    Used for: Checking results are fair for every skin tone

    Licence: CC BY 4.0

  • PubChem

    Being added

    U.S. National Library of Medicine (NIH)

    Used for: Matching other names of the same ingredient

    Licence: Public domain

  • Open Beauty Facts contributors

    Used for: Filling in product ingredient lists

    Licence: ODbL / DbCL

  • European Commission

    Used for: Correct INCI names and functions

    Licence: EU database (cited)

The studies behind each ingredient

Strongest evidence first. Where evidence is weak, we say so.

AdapaleneStrong evidence3▾
AHAStrong evidence3▾
Azelaic acidStrong evidence4▾
Benzoyl peroxideStrong evidence2▾
CeramidesStrong evidence3▾
Chemical sunscreenStrong evidence4▾
FragranceStrong evidence2▾
GlycerinStrong evidence2▾
Hyaluronic acidStrong evidence2▾
HydroquinoneStrong evidence3▾
Mineral sunscreenStrong evidence3▾
NiacinamideStrong evidence4▾
PanthenolStrong evidence2▾
PropanediolStrong evidence3▾
RetinolStrong evidence6▾
SPFStrong evidence4▾
Tranexamic acidStrong evidence5▾
Vitamin CStrong evidence3▾
AdenosineModerate evidence3▾
Alcohol denat.Moderate evidence1▾
AllantoinModerate evidence1▾
Alpha-arbutinModerate evidence3▾
BakuchiolModerate evidence2▾
  • Dhaliwal S et al. Br J Dermatol 2019;180(2):289-96 (44 adults, 12-wk double-blind RCT vs 0.5% retinol)
  • Chaudhuri RK, Bojanowski K. Int J Cosmet Sci 2014
Centella asiaticaModerate evidence3▾
CysteamineModerate evidence2▾
Green teaModerate evidence3▾
HoneyModerate evidence3▾
Iron oxidesModerate evidence2▾
Kojic acidModerate evidence3▾
Licorice rootModerate evidence3▾
PeptidesModerate evidence3▾
ProbioticsModerate evidence4▾
PropolisModerate evidence3▾
Salicylic acidModerate evidence3▾
Snail mucinModerate evidence3▾
  • Fabi SG et al. J Drugs Dermatol 2013;12(4):453-7 (25 women, 14-wk split-face RCT)
  • Tribo-Boixareu MJ et al. J Drugs Dermatol 2009 (open-label photoaging)
  • Mouse wound model 2021
SqualaneModerate evidence1▾
ThiamidolModerate evidence2▾
UreaModerate evidence2▾
Beta-glucanWeak evidence2▾
  • Pillai R et al. Int J Cosmet Sci 2005;27:292 (27 subjects, 8 wk, industry)
  • Du B et al. Carbohydr Polym 2014 review
Birch sapWeak evidence2▾
EGFWeak evidence3▾
HeartleafWeak evidence2▾
MugwortWeak evidence3▾
PDRNWeak evidence1▾
PHAWeak evidence3▾
Rice extractWeak evidence2▾
Tea tree oilWeak evidence3▾
Zinc PCAWeak evidence3▾