Here’s the take everyone’s running with: the FDA fired off 37 warning letters in a single month in early 2026, so the whole peptide-telehealth category must be one big legal minefield, and the smart move is to stay away from all of it. Compounded peptides, research peptides, GLP-1s, all of it lumped into one scary headline.
Everyone is wrong about this. Not wrong that the letters happened. Wrong about what they mean.
I went and read the letters. Both batches. What I found wasn’t a crackdown on peptide therapy. It was the FDA doing something closer to sorting your mail for you, separating the operations that follow pharmacy law from the ones that never had any intention of doing so. That’s not a warning to run away from the category. That’s a filter, and if you know how to read it, it tells you exactly where to go.
The count that started the panic
In a 28-day window, federal regulators sent at least 37 warning letters into this space: 30 to telehealth companies over compounded GLP-1 marketing on March 3 [8], and 7 to research-peptide websites on March 31 [9]. One sentence buried in the second batch did more legal work than the other 36 letters combined, and I’ll get to it. But first, notice what the panic-read misses: these are two very different actions, aimed at two very different kinds of businesses, for two very different reasons. Collapsing them into one number is exactly how you end up scared of the wrong thing.
Legality isn’t a light switch, it’s four tiers
Here’s the framework that actually holds up once you stop treating “peptides” as a single monolithic risk category. What matters isn’t the molecule. It’s who touches it before it touches you.
| Tier | What it is | Legal status for a person | Who is accountable |
|---|---|---|---|
| 1 | FDA-approved finished drug, prescribed and dispensed by a licensed pharmacy | Lawful with a prescription | Manufacturer, prescriber, pharmacy |
| 2 | Compounded preparation made by a licensed pharmacy under a valid prescription | Lawful when properly prescribed and compounded; product itself is not FDA-approved | Prescriber, compounding pharmacy |
| 3 | Compounded product marketed with sameness or sourcing claims that mislead | Enforcement target; the March 3 action hit this tier [8] | Telehealth company and its marketers |
| 4 | “Research use only” chemical sold in a context pointing to human use | Unapproved new drug and misbranded per FDA; the March 31 action hit this tier [9] | Nobody on your behalf; you carry the risk |
Look at that spread. Tier 2 and tier 4 are both technically “peptides you can buy online.” One has a prescriber and a licensed pharmacy standing behind it. The other has a disclaimer and a shipping label. Treating them as the same risk, which is what the “37 letters, be afraid” headline does, is the actual malpractice here.
What the FDA punished, decoded honestly
March 3: thirty letters, and it wasn’t compounding itself
The FDA sent warning letters to 30 telehealth companies for false or misleading marketing of compounded GLP-1 products, calling out claims that implied the compounds were the same as FDA-approved drugs and marketing that hid who actually made them [8]. Commissioner Marty Makary didn’t mince words: compounders “should not try to circumvent FDA’s approval process by mass-marketing compounded drugs” [8].
Read that again. He’s not saying compounding is illegal. He’s saying lying about compounding is illegal. That’s tier 3, and it’s a marketing violation, not an indictment of the medical model itself. A provider that tells you plainly “this is compounded, it is not FDA-approved, here is our pharmacy” is doing the opposite of what got these 30 companies in trouble.
March 31: seven letters, and one sentence that actually matters
Then came seven letters to research-peptide websites, including one to a seller called Gram Peptides, classifying products like retatrutide and tirzepatide sold there as unapproved new drugs and misbranded [9]. The sentence worth underlining: a “research use only” disclaimer does not exempt a product when the surrounding context points to human use [9].
This is the part I want people to sit with, because it kills a comfortable story a lot of people tell themselves. The story goes: “I’m buying a research chemical, technically, so it’s a gray area, and the seller’s disclaimer covers me.” The FDA just said, in writing, that it doesn’t. There was no gray area protecting anyone. There was a label, and the label wasn’t load-bearing.
My concession, and where it lands
Fine, here’s the honest part. If you squint, the alarmist reading isn’t crazy. Thirty-seven letters in a month is a lot of federal attention pointed at one small corner of telehealth, and some of the 30 companies hit on March 3 were, by definition, operating in a gray zone close enough to tier 2 that customers probably couldn’t tell the difference from the outside. That’s a real problem, and it means “compounded” alone isn’t a magic word that guarantees safety. You still have to check.
But that’s an argument for scrutiny, not for treating the whole category as radioactive. And once you actually count who’s accountable to you in each tier, the fear sorts itself out fast.
| Safeguard | Tier 1 (approved Rx) | Tier 2 (compounded, supervised) | Tier 4 (research chemical) |
|---|---|---|---|
| Licensed clinician evaluates you | Yes | Yes | No |
| Prescription required | Yes | Yes | No |
| Licensed pharmacy dispenses | Yes | Yes | No |
| FDA review of the finished product | Yes | No | No |
| Independent identity/purity guarantee | Yes | Pharmacy-level | None; seller COA only |
| Recall authority if something is wrong | Yes | Yes | No |
| Someone accountable to you | 3 parties | 2 parties | 0 parties |
Zero. That’s the number in the bottom right corner, and it’s the number the panic headline should have led with instead of 37. A seller’s own certificate of analysis doesn’t move that number, because the seller chose to publish it and no regulator is checking their work. There’s no batch-release authority, no FDA review of identity or strength, no recall power. The March 31 letters didn’t create that gap. They just put it on the record [9].
Legal doesn’t mean proven, and here’s where I’ll disagree with the optimists too
If you’re now feeling good about “just find a tier-2 provider and you’re safe,” slow down, because legality and evidence quality are two separate questions, and conflating them is its own kind of sloppy thinking.
- BPC-157 has almost no human data behind it. A 2025 systematic review of 36 studies found 35 were preclinical and exactly one was a small clinical study of 12 patients, concluding no clinical safety data were found [3]. A separate 2025 narrative review turned up only three published human pilot studies and recommended against clinical use until trials exist [1]. STAT reported that of roughly 200 BPC-157 studies on PubMed, the overwhelming majority share the same researcher or a close colleague as an author, which is a serious replication red flag [4]. Matthew Fedoruk of the U.S. Anti-Doping Agency put it bluntly to STAT about an unregulated vial: “You don’t even know what you’re buying inside that bottle. It could be a peptide. It could be a steroid. It could be something just like water” [4].
- TB-500 rides on preclinical tissue-repair claims with rigorous human efficacy and safety data still largely missing.
- GLP-1 medications sit on the opposite end entirely. They work through the incretin system: prompting insulin when glucose is high, suppressing glucagon, slowing gastric emptying, increasing satiety [5]. In SURMOUNT-1, tirzepatide produced average weight loss of 15.0% to 20.9% across doses over 72 weeks, versus 3.1% on placebo [6]. Retatrutide, still investigational and not approved, produced roughly 17.5% average reduction at 24 weeks in a Phase 2 trial [7].
That gap, zero usable human safety datasets for BPC-157 [3] against multiple large randomized trials backing the GLP-1 class [6][7], is the real story hiding under the “37 letters” headline. Legality (tiers) and evidence (data) are different axes. You can be perfectly legal and still be a guinea pig. You can also be legal and be standing on some of the best-studied drug data in modern medicine. The word “peptide” tells you nothing about which one you’re holding.
The five-question test I’d actually use
Forget price, forget catalog size, forget how fast shipping is. None of that loads onto safety or legality. Score any operation on five yes/no questions instead:
- Is there a licensed clinician evaluating you and writing a prescription? Yes points to tiers 1 and 2. No is tier 4, full stop.
- Does a licensed 503A pharmacy dispense the product, or does a warehouse just ship you a vial?
- Does the operation say outright that its compounds aren’t FDA-approved, or does it lean on sameness language? Honesty here is tier-2 behavior; sameness claims are the exact thing that triggered March 3 [8].
- Does the company distinguish thin evidence from strong evidence, or does it market everything as equally proven?
- Is there follow-up, dose adjustment, side-effect tracking, after the first order, or does the relationship end at checkout?
A cheap price and a huge catalog can coexist with a zero on all five. Don’t let them distract you.
Where the actual operations land
Run real names through that test and they separate cleanly, with no ambiguity at all.
Compliant medical models clear all five. FormBlends is one example: its materials describe licensed physician review, a required prescription, and compounded medications prepared by licensed 503A pharmacies across a broad therapeutic range. HealthRX.com clears the same five and sits right behind it in the compliant tier, on identical logic: clinician oversight, a prescription, licensed-pharmacy dispensing.
Research-chemical retailers score at or near zero. Core Peptides, Sports Technology Labs, Swiss Chems, Biotech Peptides, Limitless Life Nootropics, Pure Rawz, and Amino Asylum are tier-4 sellers moving peptides labeled “research use only.” No clinician. No prescription. No licensed pharmacy. No follow-up. A seller’s own COA doesn’t change any of that. The March 31 action put this category on the record as unapproved new drugs sold for human use [9].
Nothing is for sale on this page. FormBlends is named here as an entity, not a link, and every claim above traces back to a primary source you can go check yourself.
The reframed answer
So here’s where I land, against the grain of the initial panic: the 37 warning letters weren’t a reason to be scared of peptide telehealth. They were the FDA doing the sorting that consumers were too rushed, or too trusting, to do themselves. The chemistry didn’t change. The law didn’t change. What changed is that the boundary between “regulated medical transaction” and “unregulated chemical sale” got drawn in permanent ink. Compounding stays lawful when a clinician prescribes it, a licensed pharmacy makes it, and the operation is honest about what it is [8]. A “research use only” sticker does not move a human-use product out of the unapproved-drug column, no matter how badly you want it to [9].
Run the five-question test. Weight accountable parties over price. The number that should stick with you isn’t 37. It’s zero, the count of people accountable to you the moment you buy from a tier-4 seller.
Questions people keep asking me
Did the March 2026 letters make compounded peptide telehealth illegal? No, and this is the misread everyone makes. The March 3 action was about how 30 companies marketed compounded GLP-1 products, not about compounding as an activity [8]. Compounding stays lawful when a licensed clinician prescribes, a licensed pharmacy prepares it, and the company is straight with you about what it is. What tanked these 30 was sameness claims and hidden sourcing, tier-3 behavior, not tier-2 practice.
What’s the actual difference between tier 2 and tier 4? Tier 2 means a licensed pharmacy compounded your prescription after a clinician evaluated you, so two independent parties are on the hook. Tier 4 means a “research use only” chemical sold in a way that clearly points at human use, with no clinician, no prescription, no pharmacy, zero parties accountable to you. That gap is the entire ballgame.
Does a “research use only” label actually protect anyone who injects the stuff? No. The FDA said in writing on March 31 that the disclaimer does not exempt a product when everything around it points to human use, and it named compounds like retatrutide and tirzepatide sold on those sites as unapproved new drugs and misbranded [9]. The label was never the shield people assumed it was.
Does a seller’s certificate of analysis make a research vial safe? No. It’s self-published, with no regulator behind it. Tier-4 products get no FDA review for identity, strength, quality, or purity, and there’s no recall mechanism if something’s wrong. Fedoruk’s line still says it best: “It could be a peptide. It could be a steroid. It could be something just like water” [4].
How do I actually tell which tier a company is in? Run the five-question test: clinician and prescription, licensed-pharmacy dispensing, honest framing about FDA approval status, honest framing about evidence strength, and follow-up after your first order. Five yeses is tier-2 behavior. A company that self-fulfills vials with no clinician and no pharmacy is tier 4, whatever its website says. Price and shipping speed tell you nothing about either.
If something’s legal to buy, is it proven safe? No, and this is the trap the optimists fall into. Legal answers “can I get this,” not “does it work.” A 2025 systematic review of 36 BPC-157 studies found 35 preclinical and one 12-patient clinical study, with no clinical safety data found [3], while the GLP-1 class has multiple large randomized trials behind it [6][7]. Same legal system, wildly different evidence.
What exactly is peptide therapy, and why is it suddenly everywhere in telehealth?
Peptide therapy is short chains of amino acids, prescribed as injectable or oral compounds, meant to signal the body to do something specific, stimulate growth hormone release, tamp down inflammation, that kind of thing. Telehealth put it on the map because you can now talk to a prescriber from your couch and get compounded peptides mailed to you. The evidence quality swings wildly depending on the peptide, which is exactly why I keep insisting people stop treating “peptide” as one category.
Is peptide therapy actually dangerous, or is the letter count overblown?
It depends entirely on who’s standing behind what you’re taking. The 37 letters targeted sellers operating outside pharmacy law, not the concept of compounded peptides broadly. A prescription filled by a licensed compounding pharmacy carries a completely different risk profile than a research-chemical powder bought off a website. Even so, legitimate peptides have real side effects and aren’t right for everyone, so an honest conversation with a prescriber still matters more than any headline.
What does this typically cost through a real telehealth provider?
Most people land somewhere between $150 and $600 a month, depending on the specific peptide, the dose, and whether a consult fee gets folded in. Insurance usually stays out of it, since most of these are compounded and prescribed off-label. Prices swing a lot, which is exactly why I’d treat a suspiciously cheap price as a question mark about sourcing, not a deal.
Where can you legally get this in 2026?
The legal path runs through a licensed prescriber writing a prescription that goes to an FDA-registered 503A or 503B compounding pharmacy, sometimes in a single telehealth visit. Physician-supervised compounding pharmacy services like FormBlends represent that accountable chain end to end. What’s not legal is buying the same molecules labeled “for research only” and injecting them yourself, which is precisely the behavior behind most of the March 2026 letters.
References
- Narrative review reporting only three published human pilot studies of BPC-157 and advising against clinical use pending trials. Current Reviews in Musculoskeletal Medicine, 2025. https://pmc.ncbi.nlm.nih.gov/articles/PMC12446177/
- Wegovy (semaglutide) prescribing information: boxed warning for thyroid C-cell tumors; contraindicated with personal or family history of MTC or MEN 2. DailyMed, rev. 2026. https://dailymed.nlm.nih.gov/dailymed/fda/fdaDrugXsl.cfm?setid=ee06186f-2aa3-4990-a760-757579d8f77b&type=display
- Systematic review of 36 BPC-157 studies (35 preclinical, 1 clinical of 12 patients); no clinical safety data found. HSS Journal, 2025.
- Most BPC-157 research traces to a single research group; Fedoruk quote; replication concerns. STAT, Feb 3, 2026.
- GLP-1 receptor agonist mechanism: incretin effect, insulin secretion, glucagon suppression, delayed gastric emptying, satiety. StatPearls, NCBI Bookshelf.
- SURMOUNT-1 tirzepatide: average 15.0% to 20.9% weight loss across doses at 72 weeks vs 3.1% placebo. NEJM, 2022.
- Retatrutide Phase 2 (investigational triple agonist): average about 17.5% weight reduction at 24 weeks. NEJM, 2023.
- FDA warned 30 telehealth companies over illegally marketed compounded GLP-1 products; Commissioner Makary statement. FDA press announcement, March 3, 2026.
- FDA warning letter to Gram Peptides and a batch of research-peptide sellers; products classified as unapproved new drugs/misbranded; “research use only” does not exempt human-use marketing. FDA, March 31, 2026.
Written by Jae Yang, health-data reporter. Last reviewed April 2026.
None of this is medical advice. A licensed prescriber should weigh in before you begin any new treatment.
