Editorial policy

Each factual or efficacy claim is tied to a primary source: the published trial, the trial registry record, the official product label, or the regulator’s own database. Claims are cited inline, so you can jump from a sentence to the reference that supports it rather than trusting a bibliography at the bottom.

What we do not treat as a source: a seller’s product page, a press release, a forum consensus, a competitor’s article, or another summary of a study we have not opened ourselves. Secondary write-ups drift — a number gets rounded, a subgroup becomes the whole trial — and the only cure is to go back to the record.

Four sources do most of the work here, and each answers exactly one question. Reading one as though it answered a different question is the most common way peptide coverage goes wrong.

PubMed

What has been published, by whom, and when. It tells you a study exists and what it reported — not that the finding is large, durable, or in humans.

ClinicalTrials.gov

What was registered, its phase, enrollment, endpoints, and status. A registered or even completed trial is not a result; a completed trial with no posted results is a fact worth reporting in itself.

DailyMed

The labeling that has been submitted for a product. Listings are self-submitted by the labeller, so the presence of a record proves a filing was made, never that anything was approved or checked.

openFDA

Whether a drug has an FDA approval, and for what. This is the source that settles an approval question — and a no-match result is a finding we publish, not a gap we skip.

References are resolved live against the database that owns them and never written from memory. Author list, year, journal, title, and identifier are taken from the record returned by the source, so a reference cannot be a plausible-looking citation for a paper that does not exist or does not say what the sentence claims. Each citation is checked against the specific claim it is attached to, not merely against the topic.

Verification is not a one-time event. Pages are re-checked, and a re-check means re-opening the sources — a trial that has since reported, a label that has changed, a regulator action, a paper that has been retracted. A page whose date has moved has been genuinely re-read; we do not refresh a date to make old work look new.

On this subject, the most valuable sentence is often about what does not exist — no human trial, no published result, no approval. Those claims are also the easiest to overstate, so they are written to a fixed shape: what was searched, where, and when.

So we write “as of August 2026, none of the four completed phase 3 trials has posted results on the registry or in a journal” rather than “there is no evidence”. The first is a checkable statement about a search on a date. The second is an unbounded universal that starts decaying the moment it is published, and that nobody can verify or falsify.

The same discipline applies to safety: when we report that no interaction has been documented, the page says that nobody has looked, because absence of a reported harm and evidence of safety are different findings and readers deserve to be told which one they are getting.

Every compound carries two independent labels, kept apart on purpose. Conflating them is how a reader ends up assuming that something easy to buy has been tested, or that something well studied is available.

• Evidence tier — anecdotal, animal only, early human, or randomized trials. It describes how far the published human evidence goes and nothing else.
• Availability — FDA-approved, compounding restricted, or research use only. It describes what you can lawfully obtain today.

Every compound page also carries a required “catch” — the uncomfortable part, in plain language. It is a mandatory field in our content model, not an optional flourish, because a peptide with nothing to disclose does not exist, and a page that finds nothing to disclose has stopped looking.

Our tools convert numbers. They do not recommend, suggest, default to, or imply a dose, and this is the most rigid rule on the site.

Concretely: every input field ships empty, there are no preset buttons, there is no “typical dose” placeholder, and the dose used in a calculation comes from your keystrokes and nowhere else. The output is a conversion of what you supplied into concentration and syringe units, plus warnings when a figure exceeds the vial or the syringe. If a tool cannot compute without inventing a number, it computes nothing.

Where trial doses appear in prose, they are reported as what a study administered under supervision — a description of the past, never a schedule. See the medical disclaimer.

Content is written and edited by Grant Delaney, Research Editor. The competence claimed is specific: reading trial literature and reporting accurately what it does and does not show.

What this site does not have, stated plainly because plenty of sites in this niche imply otherwise: no medical review board, no named clinical reviewer, no peer review, no laboratory, and no product testing. If a qualified clinical reviewer is ever contracted, they will be named on the pages they reviewed and this paragraph will change. Until then, treat this as careful journalism about research, not as clinical review.

Peptideworth is funded by affiliate commission, and that funding is walled off from the evidence. Commission does not decide which compounds get covered, what evidence tier a compound receives, how favorably its trial data is characterized, or whether its “catch” gets softened. No company reviews, previews, or approves anything before publication, and unflattering findings about a company we are paid by stay in.

One rule is specific to this subject: pages about unapproved compounds carry no affiliate link and name no vendor. There is no version of this site where the honest answer is “nobody has tested this in a human” and the page next to it sells you some. The money side is set out in full on the disclosure page.

We will get things wrong. A site that never publishes a correction is either not checking or not admitting. Our corrections policy sets out what counts as an error, how to report one, what happens to a page when a cited study is retracted or superseded, and how a correction is marked so you can see it happened. Reports go to research@peptideworth.com.