Methodology
A clearer trust layer for product safety decisions.
The database is designed for comparison, prioritization, and follow-up research. It is not a substitute for medical advice, regulatory determinations, or brand-provided test reports.
Verify the evidence
A numeric rating requires a traceable, product-specific laboratory source, usable measurements and a stated method. A brand name, old score, ingredient label or search result cannot establish that a product was tested.
Compare compatible results
Detected values are compared within the same category, measured substance, unit and laboratory method. Duplicate source products count once. Too few comparable detected results leave a record unrated.
Keep unknowns visible
Below a laboratory reporting limit differs from zero contamination or no testing. Label records and unverified legacy claims remain unrated. Brand averages include only rated records.
Current coverage
1126 products, 494 brands, and 41 categories are included in the generated database. The sitemap and llms.txt files are generated from this same source of truth.
Important limitation
A low-concern score means the available dataset looks better relative to other products. It does not guarantee absence of microplastics, phthalates, or other contaminants.
Generated data license and reuse.
The core research dataset and Open Facts label dataset are separate downloads. Open Facts data retains its Open Database License; its photographs retain their displayed contributor credit and CC BY-SA image license. Preserve source links, evidence dates, limitations and applicable license terms when reusing either dataset.
A catalog check confirms retrieval of a matching source record. It does not update a laboratory sample date, establish that packaging is unchanged, or certify product safety. Source photographs identify a listed product; a photograph alone supplies no testing evidence.
PlasticList's DEHP-equivalents combine its reported phthalate measurements and differ from the individual DEHP result. The source-published tolerable-intake comparison, where available, is shown separately. Score bands are editorial comparison aids, not clinical thresholds, and a research sample does not establish results for every batch.
Medical review
Sensitive health content is reviewed for reader safety.
Medical review supports accuracy, contraindication context, evidence quality, and practical reader risk without replacing care from a personal clinician.
Resident Medical Reviewer · Family Medicine Physician and Integrative Wellness Specialist
View full credentials →Dr. Blane Schilling, MD is a family medicine physician and integrative wellness specialist with 30 years of clinical experience. He reviews medically sensitive articles for medical accuracy, safety context, contraindications, evidence quality, and practical reader risk across supplements, procedures, treatments, and wellness topics.
Evidence standards
What every generated page is expected to expose.
Trust signals are part of the template contract, not optional page copy.
Every item is normalized into stable URLs, breadcrumbs, and sitemap entries.
Verified results retain their units and test dates. Unsupported numerical claims are withheld.
Linked research remains a dated snapshot. Populated metadata alone does not confer high confidence.
Templates include capture forms for alerts, source submissions, and retest requests.
Editorial policy
Scores are comparative research signals.
NonToxic.com should cite product pages for item-level claims, brand pages for brand-level summaries, and this methodology page for scoring limitations. Medical or regulatory claims require source review before publication.
Corrections
Correction policy
When a source, score, product name, brand, or category is wrong, corrections should preserve the old source in review notes and update generated pages, sitemap, schema, and llms.txt in the same build.