cite or don't claim.
most public peptide writing is marketing dressed as science. every claim here points at PubMed, the FDA, or a trial registry, and every PMID is checked against NCBI before a commit ships.
in plain terms: this is how every peptide on reptides gets its grade, and how every claim gets its source.
evidence over opinion. citation over claim. published over promised.
the tier system.
S / A / B / C / D / F is the overall Reptides grade. Evidence confidence and grades for individual outcomes are shown separately. Overall grades shift when new evidence lands. Foundayo's FDA approval in April 2026 moved orforglipron up the GLP-1 board overnight. REDEFINE-4 missing its primary endpoint did the opposite for CagriSema.
The highest-confidence evidence record: regulator-reviewed human pharmacology or outcomes, large confirmatory trials, or replicated direct human data tied to a clear identity. An S grade can support a benefit, harm, discontinuation, or withdrawal conclusion; it is never a safety endorsement.
- regulator-reviewed human evidence or large confirmatory trials
- direct human pharmacology or outcomes tied to the exact identity and formulation
- primary sources are auditable
- approval, availability, and risk are displayed separately from the evidence grade
A strong overall Reptides grade. The letter alone does not state a trial stage or regulatory status: it can coexist with primarily preclinical evidence, while evidence confidence and grades for individual outcomes are shown separately.
- overall grade explained on the compound page
- evidence confidence shown separately
- individual outcome grades shown separately
Decent evidence, mostly preclinical with smaller clinical trials. Usable claims, depth of human data not there yet.
- ≥ 1 phase 2 trial or strong observational cohort
- consistent preclinical mechanism
- no major safety red flags
Thin or inconsistent evidence. Interesting mechanism, real-world use, not enough rigorous human data to support the marketing claims.
- preclinical real, human data small / underpowered
- popular without published validation
- plausible but unproven
Minimal human evidence. Usually fragments or derivatives whose claims significantly outrun the data. AOD-9604, for instance, is the 176-191 tail of HGH sold as a fat-loss drug with no positive phase 3 result.
- few or no human trials
- marketing claims exceed published data
- animal mechanism didn't translate
Documented safety concerns or claims with effectively zero rigorous human evidence. follistatin-344 sold as untested gene therapy, IGF-1 LR3 with tumorigenicity questions. The published safety record outweighs the evidence of benefit.
- safety signals (tumorigenicity, hypoglycemia, untested gene-therapy form)
- claims with effectively zero rigorous human evidence
how citations work.
Every substantive claim is footnoted to a peer-reviewed paper indexed on PubMed, an FDA document (approval letter, label, FAERS report), a ClinicalTrials.gov registry entry, or a sponsor SEC filing. When multiple sources exist, the highest tier wins:
Press releases without published data, blog posts, social media, supplier marketing, unsourced claims from other peptide sites, and AI-generated summaries. When only a press release exists, the citation reads “company X press release, date” and the page flags that the peer-reviewed paper is pending.
how uncertainty is handled.
Semaglutide has 17,604 patients of cardiovascular outcome data. BPC-157 has 200+ rodent papers from one lab in Zagreb and zero phase 3 RCTs. Those two molecules cannot be ranked by the same yardstick.
- Evidence strength informs the rationale. Phase 3 RCT beats phase 2 beats preclinical beats anecdotal as an evidence ladder. The page still shows the overall grade, evidence confidence, and any grade for an individual outcome separately.
- Weak evidence is surfaced, not buried. When a peptide has only preclinical data, the page says preclinical only. When 95% of the literature comes from one lab, the page says single-lab dominance and names the lab.
- No claim outruns its evidence. If a supplier pitches peptide X as reversing aging and the only paper is a 2-week C. elegans study, the page reports it as a mechanism signal in a worm, not a longevity claim in humans.
- Borderline placements are explained. When a tier could go either way, the rationale paragraph spells out what readout or regulatory event would move it.
A certificate grade measures the evidence printed on that certificate. A mass spectrum, m/z assignment, or molecular weight is the strongest identity basis. A named non-mass method such as HPLC retention-time matching is weaker, but it is still real identity testing and creates no automatic letter ceiling.
- Mass data shown. No identity ceiling; S or A remains reachable if the rest of the certificate earns it.
- Named identity method, no mass data. No identity ceiling. Retention-time matching is the weakest identity pillar, not no test.
- Bare ‘Identity: Pass,’ no method. C ceiling, because the conclusion cannot be independently inspected.
- No identity result. D ceiling. A high purity percentage cannot replace identity.
A ceiling is not an automatic grade; the other evidence can still move the letter lower. The rule is identical for a PDF and a photograph of the same certificate.
The final certificate grade uses five factors: identity (3 points), purity plus quantitation (3), contamination testing (2), net content (1), and traceability (1). The letter is then held to the strongest ceiling the printed evidence supports.
- Independent verification and in-house testing are not the same. A lab-controlled source match can reach S. An independent report without that match can reach A. A seller testing its own product cannot exceed C, and the result says so plainly.
- Purity needs real quantitative context. Purity alone cannot exceed B. A separate consistent assay earns full credit; so can a named peptide-quantitation method with three label-consistent, low-variation sample results. Gross weight, one quantity row, or arbitrary repeats do not qualify.
- Core safety evidence is not the same as optional screening. Endotoxin plus USP <71> sterility is the full shape. Endotoxin plus microbial PCR is A-eligible but weaker and cannot support S. One core result caps at B. Passing heavy-metals or fentanyl screens add no points, though paying for them shows broader testing; an explicit failure forces F.
- Every grade explains itself. The result names what the certificate did well and the specific evidence that would move it higher. Accusatory language is reserved for a concrete conflict with a lab-controlled source.
how citations stay clean.
Every PMID on the site is resolved against NCBI before a commit can land. The pre-commit hook runs scripts/audit-citations.mjs on every data-file change. If a PMID does not return a paper whose title matches the cited claim, the commit is blocked:
PMID attached to claim.
Every footnote points to a peer-reviewed paper, FDA doc, or trial registry. The abstract is pasted in before the citation gets written.
NCBI resolver runs.
Pre-commit hook hits the PubMed API for every PMID in the diff. The title must match the cited claim, not a fungal biology paper.
pass → live.
Fail → the commit is blocked, the citation never reaches main. Zero hallucinated cites in the repo.
Current count: 3,140 cited sources across the wiki, briefs, and market map, and you can read the whole list. 1,895 are peer-reviewed PubMed papers (clinical trials, cohorts, and mechanism studies), each resolved against NCBI; the other 1,245 are DOI or PMCID-only scholarly records, identity databases, FDA labels and filings, ClinicalTrials.gov registries, regulatory records, manufacturer disclosures, and lab certificates. A journal or DOI link is deduplicated when a canonical PMID for the same paper is already in the index. Found a broken one? Email reptidesinfo@gmail.com with the peptide and footnote number. Fixed within 24h, or the citation gets pulled.
how rankings change.
Tiers are not static. Recent examples of moves we made and why:
- Phase 3 readout. SURMOUNT-5 (tirz vs sema head-to-head, 20.2% vs 13.7%) reinforced tirzepatide's S tier. SURMOUNT-OSA gave it a sleep-apnea indication and a second outcome anchor.
- FDA action. Orforglipron (Foundayo) approved April 1, 2026, moved up the GLP-1 board the same day. The April 2026 503B exclusion proposal redrew the compounded GLP-1 lane and we updated every page that referenced compounded supply.
- Trial miss. REDEFINE-4 (CagriSema vs Wegovy head-to-head) missed its primary endpoint. CagriSema stayed B tier with the rationale paragraph rewritten to explain why.
- Safety / supply event. Peptide Sciences shutdown in March 2026 collapsed the largest US gray-market supplier overnight. Supply-chain pages on every affected molecule got updated within the week.
- Class expansion. Triple-agonists (retatrutide, ACHIEVE-3 oral programs) entering the field reset the ceiling for what S tier looks like in the GLP-1 class.
Every tier change is logged with a revision date on the page footer. The monthly dispatch (free, first of every month) covers the major shifts since the last issue.
community reports.
Reader up/down voting was retired. Community Trials now collects structured, anonymous self-reports and displays their sample size separately from cited research.
Community signal can prompt an investigation, but reports never enter the cited evidence shape or change a tier letter. A grade moves only when published or regulatory evidence supports the change.
conflicts of interest.
- no investors
- no sponsors
- never made a dollar from any vendor
- never received product from any vendor
- no affiliate revenue
- no display ads
- no industry consulting
- no clinic referrals
The only revenue source is reader membership. Current plans and localized prices are shown on the membership screen. Reader-funded. Run by the editor (@ogsvg on X), with @_9th_life_ as cofounder on growth and distribution; every tier ranking is the editor's. This is what lets us call GHK-Cu's tier honestly when every cosmetic supplier on the internet sells it, and keep follistatin-344 at F when the bodybuilding community pushes it as a missing link. No advertiser ever buys the verdict back.
corrections + tips.
Broken citation, outdated claim, new trial readout, factual error, missing peptide. Email reptidesinfo@gmail.com with the page URL and what is wrong. Typical turnaround is 24h. Tier changes of more than one letter (S to B, B to F) get a dispatch note so readers see the move and the reasoning.
FOUND A BROKEN CITE?FIXED WITHIN 24Hscope.
153 peptide and compound pages plus 10 outcome guides: weight loss, recovery + injury, skin + anti-aging, muscle + performance, hair growth, libido + sex, focus + cognition, growth hormone, tanning + pigmentation, longevity. Plus a market map, supply-chain map, individual claim checks, and compound-to-compound comparisons.
The board is peptide-first and intentionally includes a bounded, evidence-led set of adjacent compounds that collide with the same market and literature: metabolic research chemicals, androgen and anabolic compounds, tadalafil, and insulin. Every page states what the compound actually is; inclusion never reclassifies a steroid or small molecule as a peptide.
Out of scope: cancer therapeutics, vaccines, recreational drugs, conventional supplements, and unrelated medicines. Veterinary-only compounds can appear when that status is central to the evidence record, but animal approval is never presented as human evidence or authorization.
New pages get added when a molecule hits real community demand and has enough published evidence to write something honest about. Requests welcome at reptidesinfo@gmail.com.
RESEARCH REFERENCE, NOT MEDICAL ADVICE.