How we choose products

Price, package and date travel together.

We use reviewed public product identities and one-time package prices in the source currency. Membership, subscription, quantity and conditional promotions remain separate. Shipping and tax treatment are shown per listing; unreviewed inclusion stays unknown. Native currencies are displayed explicitly and are not converted or ranked together; the basket and price-target tools currently support USD. Newly discovered catalogs are partial source selections with unreviewed label amounts and reports. An unconditional public code can affect comparison math only when it is fresh, directly sourced and bound to the exact package; one best code is used without stacking. A code estimate is not a checkout guarantee.

Failed or blocked source checks retain the last good price and original date. History contains observed dates, not invented daily values. A changed package or CBD label interpretation starts a separate package history. Directory-only and review/disabled entries are not fabricated as offers.

CBD-only mass controls CBD unit price.

Whole-package price divided by explicitly supported whole-package CBD gives cost per mg CBD. Per-serving values require supported servings and exact package counts. Extract, formula, THC and other cannabinoid masses stay separate; an unknown CBD total has no unit price. Accessories use per-item prices. These comparisons do not establish equivalent use, absorption or clinical effect.

Vendor label interpretation records its source and review date. It is not a measured batch result. Whole-package units · Extract versus CBD

Document checks have separate scopes.

We distinguish product-family association, exact package/batch matching, issuing-lab confirmation, who acquired/submitted a sample and current inventory matching. A vendor-posted report is labelled as such. A lab-record match confirms the checked document scope; it does not mean we bought or retested a current retail bottle.

Named lab, lot, test/issue/completion dates, methods, analytes, units, reporting limits and performed panels are retained separately. ND is not zero. Missing or unperformed panels are not passes. An unavailable verification route is not proof of an invalid report. A numerical or wording mismatch is described specifically rather than converted into a safety verdict.

What lab confirmation means · Laboratory directory · Accreditation scope

Research is not a vendor score.

The paper browser combines a bounded selection of highly cited and recent PubMed-indexed studies and reviews. It is not a systematic review or exhaustive evidence map. Every imported PMID is checked against PubMed. Compound and topic tags follow titles and abstracts; study designs and model tags follow source indexing and explicit abstract terms. Citation counts and colors do not measure benefit. Short author excerpts are distinct from the smaller set of reviewed notes, which record the studied preparation, population and interpretation limits. Registration does not imply results.

Metadata can be rechecked against original records, but summaries and inclusion decisions require review. Full papers and full publisher abstracts are not republished. Locally displayed figures are checked against the published PMC article, reuse license and figure credits; each retains attribution and source links. Full-text availability alone does not grant image reuse permission. Refresh instructions and checked dates remain separate from publication dates. No retail product receives an efficacy claim because it contains a researched compound.

Make changes from evidence.

Send the preview operator a copied correction note with an exact source, affected listing or record and observed date. These content-page correction tools have no submission backend or promised response time. Price, label, report and source corrections are reviewed separately. Commercial status does not set evidence facts, source dates or default alphabetical directory order.