Our methodology

Two things are graded, separately

Most natural-health sites collapse “people have used this for a long time” and “this has been shown to work” into a single verdict. Those are different claims and we grade them on two independent axes, so a remedy can be a classical staple with almost no trial evidence, and we will say exactly that.

Evidence grade

What the published human research supports. We call it an evidence grade, not a clinical grade, because it grades literature — we are a content library, not a clinic, and this is not a clinician’s assessment of you.

A:

Supported by controlled trials in this condition.

B:

Supported by smaller or mixed clinical studies.

C:

Limited or preliminary clinical evidence.

D:

Mechanism or laboratory evidence only — no useful human data.

A grade is not a recommendation, and a higher grade does not mean a remedy will work for you. Many entries are ungraded, which means we have not assessed the literature yet — not that evidence is absent.

Traditional use

Classical:

Recorded use across multiple traditions over centuries.

Established:

A long-standing staple of at least one tradition.

Peripheral:

Used, but not a core part of any tradition's canon.

No traditional record:

Rendered as nothing at all, rather than as a claim of absence.

Traditional standing is a historical fact about practice. It is never evidence of safety or of effect, and we do not present it as either.

Citations

Where we cite a study we cite a PubMed identifier, and every identifier is resolved against PubMed before it is published — a citation that does not resolve is removed rather than reworded. We check the publication year against what our own text claims, because a real identifier attached to the wrong claim is still wrong.

Where we could not find human research, the page says so in plain words instead of describing studies in the abstract. “No human trials specific to this have been published” is a useful answer.

Safety information

Every entry carries a safety review. Pages show who should avoid it as direct instructions, and when to see a doctor with the specific symptoms that warrant it. Entries that can cause serious harm are held to a stricter standard, and for restricted or toxic substances we deliberately publish no dose, preparation method or source — the page explains the risk and points to safer options instead.

We publish pages on dangerous substances rather than omitting them. People search for them precisely because they are dangerous, and an honest page is more useful than whatever else they would find.

How entries are written and checked

Entries are drafted with AI assistance and then put through automated checks before publication: the structure and required safety fields must be complete, every citation must resolve, claims are reviewed by a second independent model for anything stronger than the evidence supports, and the writing must meet a plain-language reading standard. A draft that fails any of these is not published until it passes.

Entries involving higher-risk substances, drug interactions, pregnancy or children receive human review. Routine entries are sampled rather than individually read. We would rather state that plainly than imply a clinician has read every page.

What we do not do

• We do not give medical advice, and nothing here is a diagnosis.
• We do not claim anything cures a condition, and we do not promise outcomes.
• We do not rank remedies commercially. Where products appear, inclusion is never sold, and affiliate links are disclosed where they exist.
• We do not present traditional use as proof, or an ungraded entry as an endorsement.

Corrections

If something here is wrong, we want to know and we will fix it. Use the contact page, and tell us which entry and what is wrong with it. Corrections to safety information are treated as urgent.