Methodology

Everything we publish is built to be checkable. Here is how we work.

Sourcing tiers

• Tier 1 (preferred): peer-reviewed papers, government-agency publications (FDA, EMA, MHRA, national agencies), official regulatory documents, manufacturer technical sheets.
• Tier 2 (with context): systematic reviews and statements from credentialed experts who cite their own sources.
• Tier 3 (sparingly, attributed): community consensus and vendor lab certificates, clearly labelled, never presented as fact.

Every health, dosage, mechanism, or regulatory claim is tied to a source. We use compliance-safe framing ("the literature reports"), never prescriptive dosing.

How we verify regulatory status

Per-country status is verified against primary sources (EMA, FDA, MHRA, the national medicines agencies, and WADA where relevant) using a research → independent adversarial cross-check → human-adjudication process: one pass gathers the evidence with verbatim quotes, a second independent pass tries to disprove each status, and a person reconciles them before anything is treated as settled. Each regulatory page carries a prominent "last verified" date and is re-checked quarterly.

How we score peptide evidence

Every peptide page shows three confidence scores at the top: Evidence depth, Source recency, and Safety profile clarity. Scores are hand-graded by our editors against the rubric below, not generated by a model. They are re-graded quarterly alongside the regulatory verification cycle.

Evidence depth (0–100)

A composite of human RCT count (weighted most heavily), combined sample size across those trials, independent replication, and length of post-publication follow-up. For peptides without a meaningful human-trial base, weight shifts onto controlled animal studies and mechanism work, with a hard cap below 30. For approved medicines with millions of patient-years of post-marketing data, scores live in the 80–95 band.

Source recency (% < 3 years)

The percentage of the page's cited primary sources that were published or substantially updated in the last three years. High recency means the literature is still moving and worth following actively; low recency means the field is mature, or stalled, or both.

Safety profile clarity (0–100)

How well the adverse-event profile is characterised, not how safe the compound is. A high score means boxed warnings exist, large prospective trials describe the AE picture, and decades of post-marketing data confirm it. A low score means human safety data is sparse and most reports are anecdotal.

How we estimate the literature mix

Below the “What we don't yet know” section, each peptide page shows a stacked bar labelled Where the literature comes from. It is our editorial estimate of where the available evidence for that peptide sits, split across these categories:

• Human RCTs: randomised controlled trials in people.
• Systematic reviews: meta-analyses and Cochrane-style syntheses that aggregate prior RCTs.
• Preclinical / animal: in-vivo work in rodents or other species, plus in-vitro cell-culture studies.
• Mechanism / pharmacology: studies of how the molecule binds, what receptors it hits, and downstream signalling.
• Case reports / off-label: small observational reports, off-label use series, individual patient write-ups.
• Regulatory / agency: EMA, FDA, MHRA, WADA, national-agency documents and warnings.

These percentages are deliberately rough, because we are estimating what proportion of the existing literature sits in each bucket, not just counting what this page cites. The point is to make the shape of the evidence visible: an approved medicine with a 60% RCT slice looks very different from a research peptide whose evidence is 75% rodent work, and a reader should be able to see that at a glance.

AI and human roles

AI tools may assist drafting and research, but AI does not publish: a human editor reviews every claim against its source, and the human is the author of record. Regulatory pages additionally require a human to verify against the named regulator before publication.