Marketing lies.
Ingredients don’t.
Sketchy Labels is an independent company that translates the back of the bottle into plain English — backed by public regulatory data, not by brand partnerships, affiliate links, or vibes.
The “clean beauty” industry is a $14B category built on a 2004 paper that didn’t have a control group. Influencers tell you to fear methylparaben while selling you methylisothiazolinone. The same brand sells “tox-free” lotion and a fragranced version of the same thing.
We didn’t build a scoring system to scare anyone. We built it because the existing scoring systems were trying to.
Sketchy Labels reads ingredient lists and tells you what’s actually in them — using USDA, EU REACH and NIH PubChem data. The score is a heuristic. It’s not a doctor, a regulator, or a substitute for either. It’s a starting point for a smarter conversation about what you’re putting in or on your body.
What we’ll do.
What we won’t.
A short list of editorial principles. We violate them, you should call us out — and the contact form is right at the bottom of this page.
Cite our sources.
Every score links to the underlying regulatory or peer-reviewed data. If you want to argue with the score, you can argue with the data first.
Will doShow the math.
Our scoring methodology is published, version-controlled, and dated. You can read exactly how a 63 became a 63.
Will doUpdate when we're wrong.
Science changes. So do scores. Every guide carries a 'last updated' stamp, and corrections are logged publicly.
Will doTry it on the web.
Scan a label in your browser, no account needed. The full experience — score history, personalised results, and on-the-go scanning — lives in the app.
Will doTake money from brands.
No affiliate links. No 'sponsored scores.' No 'promoted clean swaps.' If a brand wants to be on the site, they can be on the site like every other brand — by being scanned.
Won't doFake a score.
The score is what the data says. We don't adjust it for aesthetics, what's trending, or how a brand feels about the result.
Won't doPretend to be your doctor.
The score is educational. Allergies, pregnancy, infant skin, autoimmune conditions — talk to an actual professional. We'll happily say 'we don't know' out loud.
Won't doSell your data.
What you scanned is your business. We don't aggregate, sell, or share it with brands. The Privacy Policy is shorter than this page on purpose.
Won't doWhere the data comes from.
Three public, regulator-grade sources. No “proprietary AI database,” no “clean beauty consortium,” no in-house panel of paid bloggers.
USDA
Food safety classifications, nutrient data, and agricultural substance assessments used to benchmark ingredient quality.
EU REACH
The most complete public chemical-safety registry on the planet. Restricted substances, substances of very high concern (SVHC) list.
NIH PubChem
Compound-level data: structure, hazards, peer-reviewed toxicology, CAS numbers. Used for ingredient matching across all sources.
How a score becomes a score.
Every ingredient gets weighted by concentration, regulatory status, and peer-reviewed risk data. Then we compress the result into one of five plain-English tiers.
Have a tip? Caught us wrong?
We’d genuinely rather hear it. Corrections, source suggestions, and ingredient requests are all welcome.