Search any peptide and you will find the same paragraph rewritten a hundred times: a confident summary of benefits, a mechanism stated as settled fact, and a dosing table. The tone does not change between a molecule with three Phase 3 trials and a molecule with one uncontrolled rodent study from 1994. That flattening is the single biggest failure of peptide information online, and it is what this library exists to correct.
Every profile carries a tier badge before you read anything else. The tier describes the strength of human evidence — nothing more. It is not a safety rating, not a recommendation, and not a ranking of how interesting the biology is. Some of the most scientifically fascinating molecules here sit at Tier 4 or 5 precisely because nobody has done the work.
Nine entries in this library are not peptides, and we label every one of them at the top of its profile. NAD+ is a dinucleotide coenzyme. NMN is a nucleotide. MK-677 is a spiropiperidine small molecule. 5-Amino-1MQ is a quinolinium. HCG is a glycoprotein hormone, somatropin a 191-residue protein, follistatin a 344-residue glycoprotein. Cerebrolysin and thymalin are not single molecules at all — they are animal-derived mixtures of variable composition.
We include them because they are sold by peptide vendors, filed under peptides, and discussed as peptides, and pretending otherwise would make the library less useful rather than more accurate. But the label stays on. If a vendor cannot tell you whether the thing they are selling is a peptide, that is a fact about the vendor.
We report the doses used in clinical trials and on FDA labels, as a matter of record — cited, and framed as what a study or label actually administered. Where a compound has never been tested in humans, the profile says exactly that, because “no dose has ever been established” is itself the most useful dosing fact we can give you.
We do not convert any of that into a personal protocol. We will not calculate a dose for your body weight, tell you what to inject, or write reconstitution and administration steps. That is the line, and it is a real one: reporting what a trial did is publishing; generating a personalised injection instruction is closer to practising medicine, and it manufactures false precision for compounds where no therapeutic window in humans exists. Three reasons underpin it, in order of importance.
For unapproved substances, there is no personal dose to honestly calculate. A weight-based number implies a therapeutic window — an amount shown to produce a defined benefit at acceptable risk in humans. For most compounds here that study has never been run, so a “0.3 mg for your weight” figure would be interpolated from nothing and handed to one person as instruction. We will tell you what a trial used; we will not pretend a personalised protocol exists where it does not.
For approved drugs, dosing is clinical work. It is individualised, titrated, and monitored by a physician who can see the patient. A web page cannot do that and should not pretend to.
And the regulatory reality is unambiguous. In March and April 2026 the FDA issued a coordinated set of warning letters to online peptide sellers. Every one of those sites carried “research use only” and “not for human consumption” disclaimers. The FDA's position was that the disclaimers were irrelevant, because the surrounding page copy described human effects and use. The website content was the evidence that the products were unapproved new drugs. Any site that pairs a peptide with a protocol has, in the agency's reading, marketed a drug — whatever the footer says.
A disclaimer does not neutralise the page it sits on. We would rather publish less and have all of it be true.
Profiles are built from primary literature, FDA public records including the 503A bulk substances lists and advisory committee materials, clinical trial registries, and WADA's prohibited list. Regulatory status in this field moves quickly — BPC-157 has changed category twice since 2023 — so every profile carries a date and an instruction to verify independently. When we do not know something, the profile says so.