How VialFacts checks a peptide claim
VialFacts checks peptide evidence one claim at a time. A sentence stays only when an opened primary source supports its exact wording. Numeric claims stay beside the identifier, population, sample, and measured result. Trial funders stay beside trial results. Forum accounts are quoted and labelled. Absences name the search and date. Uncertainty says why a question is unsettled. Regulatory status, evidence tier, and supply evidence remain separate. These method commitments were verified 22 August 2026.
What counts as a source?
The source must be the document that can support the claim. For a research result, that is the study record or paper. For an agency status, it is the agency's own record. For a registered trial, it is the registry entry.
A search engine, summary, news report, research assistant, or another website can locate a source. It cannot replace one. Our own notes cannot replace one either.
The accessible part of a source sets the claim boundary. If only an abstract can be opened, the page does not claim to have read the full paper. If a document cannot be opened, its unseen contents are not completed from a summary.
How is a research result recorded?
A research sentence names the compound, population, study size, measured outcome, and primary identifier. A numeric result includes its interval when the source reports one. The disclosed funder stays beside the result. If the paper discloses no funder, the page says that.
Measured positive results lead when the source reports them. Adverse findings and study limits stay in the same evidence frame. A positive result is not turned into a promise. A null result is not turned into a verdict about every possible use.
The page describes the study design instead of replacing it with an editorial verdict. A small uncontrolled report remains a small uncontrolled report. That description carries the limit without adding an editorial label.
How is an absence checked?
An absence is a claim about a defined search. The page names the compound, aliases, database, query boundary, and search date. Each spelling or development code is searched as itself.
A zero belongs to the exact query that returned it. It does not prove that no record exists under another name. The page says no published human result was found in that search rather than turning a search result into a claim about the whole world.
The same rule applies when a source is unreachable. The fact is that the document could not be opened. That fact does not establish what the document contains.
How are counts and rates handled?
Every rate travels with its numerator and denominator. A sample result names the number tested. A trial result names the number in the relevant population or arm.
Participant counts from separate papers are not added until overlap has been checked. A paper that reviews earlier cases may contain the same people. Counting them twice would overstate the number of unique participants.
A report count is not an incidence rate. A set of case reports stays a set of case reports. A spontaneous-report count stays a report count unless a defined population supplies the denominator.
How is funding reported?
Funding stays beside the result. A company-funded trial is still reported. The company is named where the paper discloses it.
Funding does not replace the result. It does not erase the result either. The reader receives the finding and the disclosed interest in the same place.
Where the source names no funder, the page states that source-level absence. It does not convert a missing disclosure into a claim that no money existed.
How is forum evidence handled?
The original thread must open. The page quotes a substantial passage and links the complete thread. The account is labelled self-reported and self-selected.
Forum evidence is not paraphrased into a verdict. A page does not claim that most people report an effect unless somebody counted. When a count exists, the page names who counted and how many accounts were included.
A forum account and a controlled study remain different evidence objects. Neither loses its identity when both appear on one page.
How do vial tests fit?
A vial test answers a product question. A clinical trial answers a human-outcome question. One does not stand in for the other.
The page states what was tested and which method was used. Identity, amount, purity, endotoxin, and sterility are separate measurements. A result from one test is not carried into a field that the test did not measure.
The guide to what is in the vial explains those distinctions in detail.
How are status, evidence, and supply kept separate?
Regulatory status records what an agency has done. Evidence records what research measured. Supply evidence records what was found in a product or sample.
An agency record does not create a trial result. A trial result does not create an agency status. A vial test does not establish a clinical outcome.
The compound register keeps those fields separate for each named substance.
Where does the verification date belong?
The date belongs to the claim that was checked. It is not a page-level substitute for older facts elsewhere on the page.
When a source is opened again, only the claims checked against it receive the new date. A layout date does not renew a claim by itself.
What happens when the answer is unclear?
The page says unclear and gives the reason. The reason may be a missing result, an inaccessible document, conflicting records, or a search that did not settle the question.
Uncertainty is not hidden by a softer sentence. An unsupported claim is cut. A supported limit stays beside the claim it limits.
What happens when VialFacts is wrong?
The page is corrected in the text. A dated entry records what was wrong and what changed in the corrections record.
A failed verification is a fact about the claim's support. It is recorded rather than quietly removed from the history.
Where can the method be checked in practice?
The About page states the mission and independence boundary. The page on what peptides are shows claim-level sources and dates on a general question. The register shows the separation between evidence and status.
Each source link is there so the reader can open the same record. The method is only as strong as that comparison.