We write down the sentence a reader actually heard. Vague promises are narrowed before evidence is judged.
PepCurrent methodology · updated July 17, 2026
We grade the claim, not the molecule.
A peptide can have strong evidence for one product, population, and outcome—and no adequate evidence for a broader online claim. That is why PepCurrent does not publish one universal score for an entire molecule.
The unit we check
Every conclusion is tied to an exact claim, claimed outcome, studied product or formulation, population, evidence type, jurisdiction, and verification date. Change one of those and the answer may change.
We identify what was studied. A familiar ingredient name does not establish formulation, sterility, strength, or equivalence.
We keep a measured biomarker separate from symptoms, function, clinical events, or an approved indication.
Human trials, uncontrolled reports, animal work, cells, and mechanisms are labeled separately. They do not trade places.
FDA staff review, committee advice, application review, approval, and compounding-list status are dated as different events.
We record what changed, which source caused it, what would change the answer again, and the prior version.
Confidence language
The label applies to the specific claim and outcome shown on the page. It is not a safety grade, buying recommendation, or score for every possible use.
EstablishedReplicated, relevant human evidence or approved labeling supports the exact use being discussed.
SupportedMore than one relevant human source, or one strong source, supports the narrow conclusion; material limits remain.
EarlyLimited or preliminary human evidence offers a signal but does not settle the outcome.
PreclinicalAnimal, cell, or mechanistic evidence exists without adequate human outcome evidence.
UnsupportedThe available record does not support the claim as stated.
UnknownAdequate evidence was not found, the product identity is unresolved, or the question has not been studied well enough.
Source and review rules
- Primary research and official records establish the factual spine.
- Social posts, podcasts, clinic pages, and vendor marketing may reveal a claim; they do not establish safety, effectiveness, or status.
- Material medical, safety, and regulatory claims receive a first-source check and an independent second-source review before indexing.
- Held drafts remain visibly labeled, carry
noindex, and stay out of the sitemap. - AI may help organize and draft. A human remains responsible for the conclusion and release decision.
Unknown is a real result. “No adequate human evidence found” is not a verdict that a compound never works. It is a boundary around what the reviewed record can support today.
The public change ledger
Priority Library and Claim Check pages show “What changed,” “What would change our answer,” and version history. The underlying structured registry connects each claim to its strongest source, product context, regulatory event, review date, and change record. Corrections follow the separate Corrections Policy.