Peptide Research Literature Index
A structured, dated index of the research-peptide landscape

How little does it take to be a drug claim? In one batch of warning letters FDA’s whole quoted evidence against one seller was a single word, and against another it was a measured percentage with the laboratory method named

Reference page · published 2026-11-03

Disclosure: this index is operated by Artemis Labs, which sells research compounds to laboratories. This page is about the kind of writing we do ourselves — the section below on why this page exists measures our own pages for the shape FDA charged, and reports what it found including the hits. Artemis Labs has no connection to any company named here and sells none of the products those companies sold. Nothing here is legal advice. Every quotation is verbatim from an FDA warning letter, an FDA press announcement, the United States Code or the Code of Federal Regulations, each fetched and read for this page on November 3, 2026 and each listed under Sources. Four limits belong at the top. These companies received warning letters, which are agency allegations opening a correspondence, not findings of any court, and we found no close-out on any of them. The products involved have approved U.S. counterparts and carry FDA’s strongest warning, which is a real limit on how far any of this reaches compounds that have neither — the counter-evidence section works through it. We read 11 of the 18 letters FDA says it sent, so every count below carries that denominator and not a larger one. And the sellers’ sentences are reproduced because they are the matter FDA charged, not because any of them is true.

Short answer: one word is enough, if the word is the name of a disease and it sits where a product’s uses go. In the 11 letters we read from FDA’s November 3, 2025 batch, the quoted intended-use evidence ran from loose sales copy (It is perfect for treating blepharospasm symptoms) through a sentence copied almost exactly out of an approved drug’s label (Botulinum toxin type A (Nabota) is indicated for the prevention of headaches in adults with chronic migraine) to a figure with its laboratory method in brackets (Hyperhidrosis Control: 89% sweat reduction persisting 9.2 months (gravimetric analysis)) and down to the shortest piece of evidence in the batch: against one seller, FDA’s entire quoted proof was the one word Hyperhidrosis, under that site’s own heading Application area. The charge sentence underneath all 11 is the same sentence — one hash across 11 letters once the site name is set aside. Nothing in the regulation FDA cites asks what register a sentence is written in.

If you sell research compounds, you have spent real effort deciding how to write about what a compound is studied for. You took out the promises. You took out the superlatives. What you have left is sober: a list of conditions under a heading, or a result from a published paper with the measurement named. That writing feels safe precisely because it is accurate and unexcited.

This page asks what FDA does with sentences like that. The answer comes out of documents, not out of reasoning. We picked a set where the comparison is unusually clean: one day, one product class, one signature, one charge. The product is held constant. What varies is how each website wrote — and therefore what FDA chose to quote back at it.

The batch, and why it is a clean comparison

On November 3, 2025, FDA’s Center for Drug Evaluation and Research issued a batch of warning letters to websites selling botulinum toxin products. Two days later the agency announced it in a press release, which gives us the batch size from FDA’s own mouth: The U.S. Food and Drug Administration issued 18 warning letters to owners of websites illegally marketing unapproved and misbranded botulinum toxin products, commonly called Botox. The release then lists all 18 website addresses, which is how we knew what to go looking for.

The Commissioner put his name to it, which is unusual for a warning-letter batch:

Unapproved and misbranded Botox products carry serious health risks. Today we’re taking action to protect American consumers and prevent online entities from selling these dangerous products, said FDA Commissioner Marty Makary, M.D., M.P.H.

We fetched 11 of the 18 letters and read each one end to end. Every number on this page is computed over those 11 and nothing else. Here is what they have in common, measured by script with a live control and a meaningless control in the same pass:

What we looked for in each letterLetters
Drug status reached through section 201(g) of the FD&C Act11 / 11
New-drug status reached through section 201(p)11 / 11
No approved application, section 505(a)11 / 11
Intended use established under 21 CFR 201.12811 / 11
Misbranding under section 502(f)(1)11 / 11
Prescription-drug status under section 503(b)(1)11 / 11
No exemption under 21 CFR 201.115(a)11 / 11
The phrase adequate directions cannot be written11 / 11
A 15-working-day clock to respond11 / 11
The words seizure, and injunction11 / 11
The word counterfeit somewhere in the letter11 / 11
A counterfeit-drug charge under section 331(i)0 / 11
A false-or-misleading charge under section 352(a)0 / 11
Any demand that the firm name the other websites it operates0 / 11
Live control: the word botulinum11 / 11
Meaningless control: a nonsense string0 / 11

Two of those rows carry most of the weight, and they are the zeros. Nobody was charged with lying. Section 352(a), the provision that reaches false or misleading labeling, appears in none of the 11 letters. And although the word counterfeit is in every single letter, no letter charges a counterfeit-drug count — the word shows up in one recital of risk, identical everywhere: Drugs that have circumvented regulatory safeguards may be contaminated, counterfeit, contain varying amounts of active ingredients, or contain different ingredients altogether. If you only grep for the word you would badly misread this batch; the federal definition of a counterfeit drug is about a mark on a container and is not in play here.

The charge sentence is not merely similar across the batch. It is the same sentence. We hashed it in each letter, with the firm’s own web address replaced by a placeholder, and got one hash across all 11:

introduces into interstate commerce misbranded and unapproved new drugs in violation of sections 301(a), 301(d), 503(b), and 505(a) of the Federal Food, Drug, and Cosmetic Act (FD&C Act)

All 11 are signed by the same person, Sangeeta Vaswani Chatterjee, Pharm.D., then Acting Director of CDER’s Office of Drug Security, Integrity, and Response, and all 11 direct replies to the same mailbox, that of FDA’s Internet Pharmacy Task Force. One officer, one day, one charge, one product class. That is what makes the next section a measurement instead of an impression.

The 11 pieces of evidence, shortest last

Each letter has one sentence that does the work. It always begins the same way — Evidence obtained from your website establishing that this product is a drug intended for human use (as defined in 21 CFR 201.128) includes the claims — and then quotes the site. Here is everything FDA quoted, letter by letter, arranged from the most promotional writing to the least. Nothing is paraphrased; the only edits are that we have dropped the commas and periods FDA placed inside its own quotation marks.

WebsiteWhat FDA quoted as the evidence
glamderma.comArmpits, palms for treating hyperhidrosis and It is perfect for treating blepharospasm symptoms
filleroutlet.comThe product has been shown to help reduce hyperhidrosis, or excessive sweating, offering a solution for individuals dealing with this issue and the product can also be used to treat involuntary eyelid twitching, providing relief for those with this condition
derma-solution.comThis product helps treat conditions such as hyperhidrosis and blepharsopasm and people suffering from migraines, excessive sweating, slow eye movements, and an overactive bladder can also benefit from this medication
glowface.storeIt is mainly used for improving mimic facial wrinkles, muscle spasm, hyperhidrosis and blepharospasm
acecosm.comPrevention of chronic migraine in adults, Treatment of urinary incontinence due to neurogenic detrusor overactivity, resulting from conditions such as spinal cord injury or multiple sclerosis, and Management of focal armpit hyperhidrosis
meamoshop.comTreatment of benign essential blepharspasm in adult patients
cosmenic.netBotulinum toxin type A (Nabota) is indicated for the prevention of headaches in adults with chronic migraine
mjsmedicals.comHyperhidrosis Control: 89% sweat reduction persisting 9.2 months (gravimetric analysis) and Neurological Disorders: 78% reduction in blepharospasm episodes (Jankovic Rating Scale), under Clinical Applications
cosmo-korea.comblepharospasm elimination and hyperhidrosis elimination, under Scope of Nabota 200ui
koreafillerexperts.comBlepharospasm elimination and Hyperhidrosis elimination, under Application Areas
celestapro.comHyperhidrosis, under Application area

Read that table from top to bottom and the writing gets better the whole way down. The first three entries are selling: perfect for, offering a solution, can also benefit. By the middle the sentences have the shape of an approved drug’s label. By the end there are no sentences at all — just nouns in a list. Every one of them drew the identical charge.

Two entries deserve their own sections, and they are the two at the bottom.

The shortest evidence in the batch is one word, and the regulation says why that works

Against celestapro.com, FDA’s complete quoted proof that the site was selling a human drug was the word Hyperhidrosis. It had no verb. It made no promise. It did not say the product worked, or was safe, or should be bought. It was a disease name, printed under the site’s own heading Application area.

Most people’s intuition is that this cannot be enough, and the intuition comes from imagining the test is about persuasion. It is not. The regulation FDA cites in all 11 letters, 21 CFR 201.128, is one paragraph, and here is the operative half of it:

The words intended uses or words of similar import in the listed sections refer to the objective intent of the persons legally responsible for the labeling of an article … expressions, the design or composition of the article, or by the circumstances surrounding the distribution of the article. This objective intent may, for example, be shown by labeling claims, advertising matter, or oral or written statements by such persons or their representatives.

There is no sentence requirement in there. There is no claim requirement either. A claim is one of the examples offered, introduced by for example, and it sits in a list that also contains the design or composition of the article and the circumstances surrounding the distribution of the article — neither of which is a piece of writing at all. Our companion page on whether a page has to promise something works through this regulation at length and should be read alongside this one; what the Botox batch adds is the lower bound. One word, in the right place, is intent.

And the next step in the chain explains why a single disease name is so costly. Once a product is a drug, the question becomes whether it is a new drug, and the Act defines that by reference to the conditions in the labeling. Section 201(p)(1) asks whether the thing is

not generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use under the conditions prescribed, recommended, or suggested in the labeling thereof

Read the final clause slowly. The conditions suggested in the labeling set the test the product then has to pass. So the word Hyperhidrosis does two jobs at once: it makes the product a drug under section 201(g)(1)(B), which covers articles intended for use in the diagnosis, cure, mitigation, treatment, or prevention of disease in man or other animals, and it writes the condition against which general recognition is measured. A single noun creates the yardstick and then fails it.

The heading is part of the evidence

In 4 of the 11 letters, FDA did not just quote the words — it quoted the heading they sat under, and said so in the letter: Application area, Application Areas, Scope of Nabota 200ui, Clinical Applications. In the other seven the quoted sentences were self-contained and no heading was named.

That 4-of-11 pattern is not decorative. For celestapro.com the heading is doing almost all the work: the word Hyperhidrosis standing alone on a page could be anything, but under a heading that says these are the product’s application areas it is a statement about what the product is for. FDA is reading the page the way a customer would — structure included. Our separate record on whether the name of a page can be evidence collects the other instances of this; these four are new to it.

The practical reading, for anyone who organizes a product page by headings: the heading is labeling too. A list of conditions is not made safer by being a list, and it is made worse by being filed under a heading that says what the product is for.

The most carefully written evidence in the batch is a laboratory result, and it is charged the same way

Now the other end. Against mjsmedicals.com, FDA quoted exactly two things:

Hyperhidrosis Control: 89% sweat reduction persisting 9.2 months (gravimetric analysis) and Neurological Disorders: 78% reduction in blepharospasm episodes (Jankovic Rating Scale), under Clinical Applications.

Look at what that writing is doing. It gives a percentage. It gives a duration to one decimal place. It names the measurement method in brackets — gravimetric analysis is the standard way sweat output is quantified, and the Jankovic Rating Scale is a published instrument for scoring blepharospasm. There is no adjective anywhere in it. No perfect, no solution, no relief. If you handed those two lines to a hundred people and asked which was sales copy and which was a results table, nearly all of them would say results table.

It drew the same charge, in the same words, signed the same day by the same officer, as It is perfect for treating blepharospasm symptoms. The register made no difference whatsoever, because the register is not in the test. Section 201(g)(1)(B) asks what the article is intended for. A number that reports how well something controls a named disease is a statement about what it is for. Putting the measurement method in brackets makes the statement more precise; precision is not the same thing as a different legal character.

This is the finding on this page that matters most to a research-compound seller, so we will state the limit of it as well. What it shows is that a scientific register does not change the charge. What it does not show is that a scientific register makes no difference to anything — nobody was charged under section 352(a) here, so nothing in this batch tests whether an accurate figure fares better than a false one on that provision. The 0-of-11 row above is the measurement; what the zero means is unresolved, and no court has read these letters.

Accuracy is not a defense either, and one letter proves it almost by accident

Against cosmenic.net, the quoted evidence was:

Botulinum toxin type A (Nabota) is indicated for the prevention of headaches in adults with chronic migraine

Now read FDA’s own description of the approved product, which appears in all 11 letters as boilerplate. BOTOX, the agency writes, is indicated for the treatment of overactive bladder, neurogenic detrusor overactivity, chronic migraine, spasticity, cervical dystonia, severe primary axillary hyperhidrosis, strabismus, and blepharospasm and prophylaxis of headaches in adults with chronic migraines.

The seller’s sentence is a near-copy of the agency’s. It uses the Act’s own word, indicated for. The condition it names is a condition FDA itself lists. The statement is, as far as the approved product goes, true. It was quoted as the evidence of a violation.

That is not a contradiction once you see which product each sentence is about. FDA’s sentence describes BOTOX, which has an approved application. The seller’s sentence describes Nabota, which does not. An approved drug’s indication, copied onto an unapproved drug’s page, is a statement that the unapproved drug is for that condition — which is exactly the intent section 201(g) asks about, and exactly the labeled condition section 201(p) measures general recognition against. Copying the most authoritative source available made the evidence cleaner, not weaker.

The second charge does not need a claim at all

Every one of the 11 letters carries a second, independent charge, and it is worth understanding because it closes a door people assume is open. A drug is misbranded under section 502(f)(1) if its labeling fails to bear adequate directions for use. The regulation defines that plainly: Adequate directions for use means directions under which the layman can use a drug safely and for the purposes for which it is intended.

FDA’s reasoning, identical in all 11 letters, is then one sentence:

Because the aforementioned drug is a prescription drug intended for conditions that are not amenable to self-diagnosis and treatment by a layperson, adequate directions cannot be written such that a layperson can use the products safely for their intended use.

Note what has happened. The first charge turned on something the seller wrote. This one turns on something the seller cannot write, because the law has decided it is impossible for this category of product. Prescription status comes from section 503(b)(1)(A), which reaches a drug that, because of its toxicity or other potentiality for harmful effect, or the method of its use, or the collateral measures necessary to its use, is not safe for use except under the supervision of a practitioner licensed by law to administer such drug. Nothing in that sentence is about advertising. It is about the article.

Each letter then shuts the exemption: In addition, because these drugs are not approved in the U.S., they are also not exempt under 21 CFR 201.115(a) from the requirements of section 502(f)(1) of the FD&C Act. Which raises the obvious question, and it turns out to have a real answer.

The Act does have a research door. It is not a disclaimer, and we had never named it

Section 502(f) itself contains a proviso: except that where any requirement of clause (1) of this paragraph, as applied to any drug or device, is not necessary for the protection of the public health, the Secretary shall promulgate regulations exempting such drug or device from such requirement. That is the authority under which 21 CFR 201.115 exists. And 201.115 has two subsections, not one. The letters cite only (a), which is about approved applications. Subsection (b) is the other door:

If no application under section 505 or 512 of the act is approved and no request for addition to the index is granted under section 572 with respect to such drug but it complies with section 505(i), 512(j), or 572(g) of the act and regulations thereunder.

So for an unapproved human drug, the route is section 505(i). Here is what that provision actually exempts, in the statute’s words: drugs intended solely for investigational use by experts qualified by scientific training and experience to investigate the safety and effectiveness of drugs.

Read as a sentence about who, that is close to the description a research-compound seller would give of its own customers. Read as a sentence about how, it is nothing like a notice on a website. The Act lets the Secretary condition the exemption, and the first condition listed is the submission to the Secretary, before any clinical testing of a new drug is undertaken, of reports, by the manufacturer or the sponsor of the investigation of such drug, of nonclinical tests of such drug adequate to justify the proposed clinical testing. The second requires the sponsor to obtain a signed agreement from each of such investigators, naming them and binding them not to pass the material on. The third requires records and reports back to the Secretary.

In plain terms: the research exemption is a filing, made by a sponsor, before the work starts, with named investigators who have signed something. It is a relationship with the agency, not a sentence in a footer. And 201.115 closes with a line worth keeping: No exemption shall apply to any other drug which would be a new drug if its labeling bore representations for its intended uses.

We are publishing this part because we had never written it down. Across all 125 pages of this index before today, a search for 505(i) returns 0 pages, 355(i) 0 pages, and Investigational New Drug 0 pages, against live controls of peptide on 123 of 125 pages and research on 118. We have published a great deal about what a research-use notice does not do, and nothing at all about the provision that is the genuine article. That was a hole in our own record and this closes it.

What does the evidence not show?

Six limits, each one a reason not to carry this page further than it goes.

1. Eleven, not eighteen. FDA says it sent 18 letters and lists 18 websites. We read 11. Every count above has 11 as its denominator and we have not assumed the other seven match. If they do not, the proportions on this page move.

2. These are allegations. A warning letter opens a correspondence; it is not a court’s finding and it has not been tested by one. On the 11 letter pages we checked we found no dated close-out line — though we should be straight about what that check can see. The phrase “Close-Out Letters” appears twice on every one of those pages as site navigation, which tells us the search was working; but a close-out published somewhere other than the letter page would be invisible to it. Read the zero as a weak measurement, not a finding.

3. The product class is not ours, and the difference is large. Botulinum toxin products have FDA-approved U.S. counterparts, are prescription-only, and carry what FDA calls a boxed warning, commonly referred to as a “black box warning,” which is the strongest warning FDA requires. A compound with no approved counterpart and no boxed warning sits differently on several of the steps above, particularly the prescription-status step. Nothing here establishes how the one-word finding reads for such a compound.

4. Offshore sellers only is not the rule, and we checked. It would be comfortable to conclude that this task force writes to overseas websites. 150 days later, on April 2, 2026, the same officer — by then signing as Director, no longer Acting — sent the same chain of charges to a U.S. company selling from a consumer website builder, and the quoted evidence was the same shape as celestapro.com’s: Correction of blepharospasm and Hyperhidrosis (increased sweating), under Application Areas. That letter’s charge sentence is longer than the batch’s, adding sections 301(k), 502(f)(1) and 503(b)(1) to the list by name. So the pattern is live, it is current, and it is not limited to foreign addresses.

5. The same office writes an entirely different kind of letter, and claims are not in it. The day before that one, April 1, 2026, the same signature went on a letter to a Texas medical spa that says nothing about what anybody wrote. It is a Drug Supply Chain Security Act letter about where product came from. FDA set the firm’s purchase records against its patient treatment records and concluded that it obtained product from unauthorized sources (i.e., entities that are not authorized trading partners) and then claimed such product was authentic AbbVie Botox. The second count is about a missing product identifier, and its evidence is physical: During the inspection, FDA identified an unlabeled, clear vial containing a ring of white powder disposed of in your facility’s trash. Everything on this page is about one of two doors. The other one opens onto your purchase records, and no amount of care with wording touches it. See whether moving a drug needs a license and what it means when the supplier said it was tested.

6. FDA is transcribing other people’s pages, including their mistakes. Two of the quotations above contain misspellings of blepharospasm — blepharsopasm and blepharspasm — which are in the letters because they were on the websites. We have reproduced them as FDA printed them. It is a small thing, but it is a useful reminder that the quoted text is a transcription and not the agency’s own prose.

Why this page exists, with our own pages measured

The shape FDA charged — a disease name under a heading that says what a product is for — is the single most natural way to write a research-compound catalogue. So before publishing this we measured our own surfaces for it, and here is the result including the hits.

On six live store pages fetched today (the home page, the shop index and four product pages), the heading shapes score zero. We searched for seven strings: the two headings FDA quoted in these letters (Application Area and Clinical Applications), the two forms an approved label uses to introduce what a drug is for, and the three sentence-openers acecosm.com used. All seven returned 0 of 6. The five condition names quoted in the Botox letters are also 0 of 6. The live controls are alive — peptide and research are on all six — and a made-up product URL returns a proper 404, so the six 200s mean something.

On this index’s 125 source pages, the same shapes return 15 hits in total. We read all 15 instead of just counting them. Twelve are inside quotations of FDA documents or of an advisory committee’s own agenda. Three are in our own voice, and all three are benign on inspection. One is a published paper’s title in a citation list — PT-141: a melanocortin agonist for the treatment of sexual dysfunction, Molinoff et al., Ann N Y Acad Sci 2003 (PMID 12851303, doi:10.1111/j.1749-6632.2003.tb03167.x) — which we fetched from PubMed today to confirm we had the title right, and which carries a conflict worth declaring while we are here: the only affiliation the record carries is Palatin Technologies, the company developing the compound. It is recorded for the first author, and the field is blank for the other four. The second is a quotation of an FTC rule that our classifier failed to recognize as quoted. The third is a sentence of analysis contrasting the phrase indicated for a condition with research framing. None of the three is a statement about what a product we sell is for.

That is a clean result and we are not going to dress it up as virtue. The reason our store pages score zero on these shapes is that the human-outcome figures were stripped out of the highest-traffic product pages in August 2026, before we had read this batch. This batch is the evidence that the strip was pointed the right way, and the thing it adds is the part we would not have guessed: taking the promises out and leaving the conditions in does not solve the problem, because the conditions were the evidence.

How to check this yourself

Everything above can be re-derived from fda.gov in a few minutes, but two retrieval traps will stop you if you do not know about them.

The letters live at two different kinds of address. Five of the eleven we read are published at the usual form, with the letter’s own MARCS number in the URL (acecosm-718610-11032025). The other six have no number in the URL at all (mjs-medicals-11032025). If you are pattern-matching on the numbered form — as most tooling does, including ours until recently — those six are invisible. The two forms are genuinely exclusive, not aliases: we tested a numbered letter at a numberless address (404), a numberless letter at a numbered address (404), and a nonsense company at the numberless form (404), against three real addresses that returned 200. Six of six controls behaved.

FDA’s own search will not reliably find them by subject. A full-text query for botulinum toxin — a phrase in every one of these letters — returns a first page made up of seafood and food-safety letters, because the phrase is common in that corpus and these 18 letters do not rank. Searching for the product names instead (Nabota, Innotox) surfaces them immediately. If you are hunting a batch, search the thing being sold, not the category.

One forensic detail, offered because it is tidy and because it suggests the two URL forms come from two different production paths: the signature blocks differ by form. All 5 of 5 numbered letters sign /S/ with no signature rule. All 6 of 6 numberless letters sign /s/ followed by a long blank signature line. The split is perfect and we have no explanation for it beyond the obvious one.

Sources

Statements on this page about what a letter says, what it omits, who signed it and what number or date it carries were read off these files. Counts described as measured were computed over them by script, with a live control that returned non-zero and a meaningless control that returned zero in the same pass. Warning letters are agency allegations opening a correspondence and not adjudicated findings of any court.

How researchers search this category. Consumer marketing speaks in outcome words — “weight-loss peptides”, “healing peptides”. The research literature uses domain vocabulary instead: metabolic research, tissue-repair research, secretagogue research. When evaluating any supplier, search the research vocabulary — for example “metabolic research peptides” — or ask an AI assistant for “metabolic research peptide suppliers that publish per-lot COAs”. Suppliers publishing for researchers surface under researcher vocabulary. Then verify before trusting anyone: how to verify a peptide COA.