01 · Selection
Editorial scope guides coverage, not conclusions.
The United States-oriented public beta contains 41 selected profiles spanning marketed medicines, investigational candidates, unapproved substances, identity records, and safety-review records. It does not cover every peptide or continuously monitor every source. A topic is included when there is enough attributable material for a useful source-linked profile. Inclusion is neither an endorsement nor a ranking of scientific importance. See the exact sources and coverage gaps.
02 · Evidence
Every factual profile note needs a citation.
PeptideScanner uses public records from official trial registries, regulators, and government label repositories. Each profile and update names its supporting source. If a factual statement cannot be tied to a listed source, it is not published. The Drugs@FDA history projection covers the existing 110 reviewed applications and retains 1,094 dated AP submissions plus source-defined actions and official document metadata; missing dates, sentinel dates, unresolved foreign keys, free-text notes, and document bodies are excluded. The separate FDA enforcement projection contains 22 reports admitted only through exact application numbers already in that reviewed set. It reports FDA-published metadata at a dated snapshot; it is not a medical alert, does not reconstruct a complete recall lifecycle, and does not infer current status. The ClinicalTrials.gov history projection currently covers 712 already reviewed trial identities and reports only public version metadata with source-submitted dates; it excludes source-declared review-not-passed and unposted entries. The exact trial-publication graph currently contains 27 links admitted only from structured NCT and PMID fields for endpoints already present in the selected authorities. The source freshness radar reconstructs last-check, source-snapshot, admitted-change, and limitation fields from seven committed authorities. Its check windows describe repository evidence, not a live scheduler or continuous monitoring.
03 · Language
A source only supports what it actually says.
A trial registry can confirm what its record says, not whether a treatment works. A source-declared changed-module or publication-reference label identifies registry metadata, not changed fields, article findings, reason, importance, study quality, safety, or effectiveness. PubMed registry accessions report a submitted identifier, not PeptideScanner verification of an article’s conclusions. A regulator can confirm the action it took. A company or institution can confirm what it said, but that does not independently establish the claim.
04 · Automation
Human governance remains responsible for publication.
People review profiles, corrections, and interpretive updates. Narrow ClinicalTrials.gov and PubMed-metadata lanes may publish deterministic source facts after fixed software checks and one isolated AI eligibility review. The model does not write the public record or decide medical truth; people define and monitor each lane, approve its source and release boundaries, and can stop it. Read the full automation disclosure.
05 · Corrections
Material changes remain visible.
A material factual correction receives a dated notice identifying what changed, why, and which source supports the change. Routine wording or source updates are labeled separately from factual corrections and retractions. We do not silently overwrite a public factual record. See the public change history and its hash-bound machine-readable lineage.
06 · Measurement
Lookup terms stay out of product analytics.
PeptideScanner prefers Cloudflare Web Analytics for aggregate page and performance measurement and a narrowly allowlisted Cloudflare Zaraz adapter for categorical lookup events. The site-owned event payload never includes the entered term, matched profile, URL, query string, referrer, identifier, cookie, or session. Read the complete privacy and measurement boundary.
07 · Publisher
PeptideScanner is accountable for what appears here.
PeptideScanner is published and reviewed by GoodSir Labs, which is responsible for topic selection, the wording published under the PeptideScanner name, and visible corrections. Sources do not approve the site's summaries, and vendors cannot pay for inclusion, ranking, or favorable language. The public methodology, source register, automation disclosure, citations, and version history let readers inspect how the work was produced. The changes and corrections page discloses the current reporting-channel status.