Find the named substance
Labels can use INCI names and synonyms. The reference uses exact matching to reduce false alerts from similar-looking names.
THE THINKING BEHIND GLOW
Good skin decisions start with evidence.
Project Glow brings a skin photo, a person's goals, published ingredient research and product-label checks into one clearer conversation. The result is cosmetic guidance with its limits made visible.
01 / A CLEARER PATHConcern → evidence → ingredient → product
IN THE CURRENT REFERENCE
A dated snapshot of the app's working data, rather than a claim that every entry has the same strength of evidence.
66skin concern entries
100ingredients in the library
98unique PubMed studies cited
2,699official EU and Canada list entries
From the internal Glow Skin Reference generated 7 October 2026. Counts and checks can change as the app is updated. See source notes ↗
01 / HOW THE ENGINE THINKS
Several signals. One explained path.
The working reference describes a condition engine that weighs different kinds of evidence before matching ingredients. A single image does not make a diagnosis.
- 01
Listen and look
A consented photo is considered with skin goals and profile details such as skin type and sensitivity.
- 02
Weigh the signals
Photo findings, marker scores and profile information contribute different levels of confidence.
- 03
Read the evidence
Potential ingredients are linked to concerns through an evidence library with visible quality grades and cautions.
- 04
Check before choosing
Safety rules and product-label screening add context before a person explores a product or speaks to a clinician.
Signals are combined and filtered; they are not treated as equally certain. The full reference explains the confidence method and safety gates.
INGREDIENTS, WITH CONTEXT02 / EVIDENCE LIBRARY
Not every ingredient has the same evidence.
The reference groups ingredients by strength of evidence: strong, moderate or emerging. A useful recommendation also explains the study type, formulation, possible irritation and who the evidence may apply to.
Strong
Higher-quality trials, reviews or guidelines.
Moderate
Smaller trials or authoritative reviews.
Emerging
Early or indirect evidence that needs more care.
In the 7 October reference, 72 of 100 ingredients had a PubMed re-check; 28 retained earlier curated citations and were marked for re-check. We show that distinction rather than treating the whole library as equally verified.
03 / LABEL & RULE CHECKS
A label deserves a closer read.
The reference screens named ingredients against imported prohibited and restricted lists from the European Union and Health Canada, then explains what a match means.
Separate prohibited from restricted
A restriction may depend on concentration, product type or conditions of use. It is not automatically a ban.
Keep the source visible
Each alert needs its jurisdiction, rule and source date so the original list can be checked.
What the check cannot establish
A label scan cannot measure an undisclosed concentration, find an unlisted contaminant or certify legal compliance. The current imported lists do not cover every country or every cosmetic annex.
SOURCE NOTES
Method and counts above are summarised from Project Glow's internal Glow Skin Reference, generated 7 October 2026. The reference records its own pending evidence checks and coverage limits.
04 / THE HUMAN PART
Clear limits make the guidance more useful.
Skin analysis can support informed skincare choices; it cannot diagnose a condition. An ingredient match is a starting point, not proof that a finished product will work for every person. Persistent symptoms or treatment decisions belong in a conversation with a qualified clinician.
See sources and limits