Disclosure · Updated May 2026
What Plassic infers, what it doesn't detect
Read this before assuming the score number means we measured the plastic in your bottle. We didn't. Here's what the score actually is.
1 — What the score is
An inference
Computed from a curated database of product compositions, ingredients, materials, packaging types, and peer-reviewed research on plastic leaching and microplastic shedding. The number on your screen is our best estimate of relative plastic-exposure risk for that product, given what the published science currently knows.
“Inference” is the precise word. We reason from known inputs—materials, manufacturing category, usage context, research citations—toward a probability-weighted risk estimate. We do not observe particles directly.
2 — What the score is not
Four things the score cannot tell you
- Not a measurement. It is NOT a laboratory measurement of microplastic particles in your specific product. No phone camera, app, or consumer-facing scanning tool currently in existence can perform particle-level detection in a product you're holding.
- Not a guarantee. It is NOT a guarantee that the product contains, or doesn't contain, microplastics. Scores are probabilistic estimates. A score of 90 means lower inferred risk, not zero risk.
- Not medical advice. It is NOT a diagnosis, treatment, or medical advice. Plassic does not assess individual health outcomes. If you have concerns about plastic-related health risk, consult a qualified clinician or toxicologist.
- Not a regulatory finding. It is NOT a regulatory determination. Plassic scores are independent research-derived estimates and carry no force of law. They do not constitute approval, recall, or safety classification under any jurisdiction.
3 — Why we're telling you this
The honesty clause
Consumer apps claiming “AI detection” of microplastics are common. Many are misleading. The phrase implies instrumental measurement that no smartphone-camera pipeline can deliver. We think that's harmful, because it erodes trust in the whole category and, more importantly, in consumers' ability to make informed decisions.
Plassic scores exposure honestly because trust is the moat. A score you understand and can interrogate is worth more than a confident number with no audit trail. This page is part of that audit trail. The methodology page is another part.
We also note that this disclosure is required by Apple App Store Review Guideline §5 (deceptive apps) and is consistent with the FTC's guidance on substantiation of objective product claims. We don't just follow those rules because we have to; we follow them because they're right.
4 — How the score is computed
The short version
Each Plassic score is a weighted blend across five signal classes: primary material composition, packaging type and contact surface area, manufacturing and processing conditions, known additive profiles (plasticisers, stabilisers, coatings), and category-level peer-reviewed evidence on leaching or shedding rates under normal use conditions. Weights are calibrated against published toxicological studies and reviewed periodically as new research is indexed.
The score runs 0–100. Lower numbers indicate higher inferred exposure risk. Higher numbers indicate lower inferred exposure risk. When our confidence in the underlying data is below threshold, the app flags it explicitly — a score without a confidence signal is incomplete.
For citations, weighting methodology, data-source inventory, and known limitations, read the full methodology. That page is the deep version of this one.
5 — What you should do with the score
How to use it responsibly
Use the score as one input among many in your purchase decisions. It sits alongside ingredient labels, third-party certifications, price, availability, and your own tolerance for risk. It is not a veto and it is not permission.
Higher scores generally indicate lower microplastic-exposure risk based on what we currently know. That caveat — currently know — is doing real work. The science of microplastic risk is active and contested. What we know in 2026 may look different in 2028.
Brand reformulations and new research can change scores. We re-score products periodically as our database is updated. If a product you scan today carries a different score in six months, that is a feature, not a bug.
When the data underlying a score is sparse or uncertain, we say so. A lower-confidence flag in the app is not a failure state; it is the system working correctly. A score presented without uncertainty acknowledgement is a score you cannot trust.
6 — Contact
Questions about this disclosure
If you are a researcher, regulator, journalist, legal professional, or user with a specific question about our methodology or the basis for a particular score, write to us. We read this inbox.
For general support, use the in-app feedback form. For press enquiries, see /press.