A vial arrives with a label, a compound name and a milligram figure. Behind it is a supply chain of two or three steps, and the steps are not interchangeable: each one adds or removes something that can be checked afterwards.
This page describes those steps, and then tests the description against a file FDA publishes — the list of every Drug Master File submitted to the agency. It is one of the few places where the question "who actually makes this" has a public, countable answer.
The three channels
Almost all synthetic peptide material begins in the same kind of place: a contract synthesis house running solid-phase peptide synthesis, overwhelmingly in China or India. That is true of the material inside an approved drug and of the material inside a research vial. The molecule's origin is rarely what separates them.
What separates them is what happens to the powder next.
The regulated channel. A company holding an approved application buys drug substance from a supplier whose manufacturing is documented to the regulator, and makes a finished product. Identity, potency, sterility, endotoxin and stability are all specified, tested and held against a standard, and a named company carries the legal responsibility for the vial.
The compounding channel. A pharmacy prepares a preparation for a patient. The two forms of compounder are defined by sections 503A and 503B of the Food, Drug, and Cosmetic Act, and the distinction between them is substantive — one is tied to individual prescriptions, the other registers as an outsourcing facility and takes on obligations closer to a manufacturer's. Both sit inside a framework with named duties, and both are limited in what bulk substances they may use at all.
The research-chemical channel. A repackager buys bulk powder, divides it into vials, lyophilises and labels it, and sells it online under a research-use-only designation. It is that stated use, rather than anything about the chemistry, that places the transaction outside the framework the other two channels sit inside.
The consequential point is that the same drum can feed all three. A buyer holding a vial from the third channel is not necessarily holding worse material. They are holding material about which far less is recorded, from a step that is not required to record it.
What the public file shows
FDA publishes its DMF list quarterly. The second-quarter 2026 release is current through DMF 044443 and covers files received by 30 June 2026. It holds 41,253 records, each with a number, an activity status, a type, a submission date, a holder and a subject.
A Type II file is the drug-substance category — in FDA's words, "Drug Substance, Drug Substance Intermediate, and Material Used in Their Preparation; or Drug Product". Counting active Type II files by subject line gives this:
| Compound | Active Type II files |
|---|---|
| Semaglutide | 95 |
| Tirzepatide | 40 |
| Liraglutide | 23 |
| Retatrutide | 11 |
| Oxytocin | 7 |
| Triptorelin | 4 |
| Chorionic gonadotropin | 3 |
| Gonadorelin | 2 |
| Thymosin alpha 1 | 2 |
| Glutathione | 2 |
| Aviptadil | 1 |
| Bremelanotide | 1 |
| Sermorelin | 1 |
| BPC-157, ipamorelin, CJC-1295, GHRP-2, GHRP-6, melanotan, tesofensine, epitalon, Selank, Semax, noopept, kisspeptin, follistatin | 0 |
The compounds with a large approved market behind them have dozens of registered suppliers. The compounds this market actually sells most of have none.
The count that would have been wrong
Searching the file for "BPC" returns nine rows. Not one of them is BPC-157.
Eight are British Pharmacopoeia grade designations on unrelated substances — permethrin BPC from Dr Reddy's, amoxicillin trihydrate BPC, chlorhexidine acetate BPC from Orion, nitrofurazone NF-BPC — and one is a Berry Global packaging material. BPC is the abbreviation for the British Pharmacopoeia Commission as well as for the peptide, and a keyword count reported without opening the rows would have published a registered BPC-157 supplier that does not exist.
The census above was read row by row for that reason. It is the kind of error that is invisible once it is in a table.
Retatrutide, and an industry preparing early
Eleven active files name retatrutide. Every one was submitted between 19 March 2025 and 30 March 2026, and every holder is a peptide manufacturer in China or India: Sinopep-Allsino, Harbin Jixianglong, Chengdu Shengnuo, Hybio, El-Peptido Shanghai, Guizhou Utide, Zhejiang Peptites, Hangzhou Thinheal, Nanjing Cellnuo, Fujian Genohope and Sichuan Jisheng.
Retatrutide has no marketing approval anywhere. Eleven manufacturers have nonetheless documented their production of it to a foreign regulator, ahead of any product existing to reference their files. The dose record for the compound is still a set of trial arm labels; the supply chain is already building.
The limitation this page will not leave out
Tesamorelin returns zero records, and tesamorelin is the active ingredient of an approved product.
That single row disciplines everything above. DMFs are voluntary — FDA states they "are not required by statute or regulation" — and their principal function is to let one company's confidential manufacturing information support another company's application. A firm holding its own approved application can put the same information inside that application and never open a DMF. The instrument is, in practice, mostly a generics instrument, which is exactly why semaglutide has 95 filings: suppliers are positioning for patent expiry.
So a zero in this file means one thing and one thing only: there is no registered US drug-substance filing under that name. It is not a finding that a compound has no manufacturer, that its manufacture is unlawful, or that nothing is known about it. Reading it as any of those would be the same category of error as reading "99% pure" as "99% peptide", which the assay page takes apart at length.
What the zero does support is narrower and still useful. For the compounds this market sells most of, no supplier has yet found it worth telling a regulator how they are made — because there is no approved product for such a file to serve.
What this changes for a vial
Nothing on a label is affected by any of this. A research vial's label is not required to name its synthesis house, and in practice does not.
What the census changes is the meaning of an appeal to provenance. "Pharmaceutical grade", "made in a GMP facility" and "from a licensed manufacturer" are phrases that appear constantly in this market, and for most of the compounds above there is no registered filing behind any of them. Where a certificate of analysis exists, it documents a test performed on a sample — which is a different and more checkable claim than one about who made the material, and it is the one worth reading.
Decisions about administering anything to a person belong with a clinician. This page is a description of a supply chain and a count of a public file, and nothing on it is a recommendation.