Are grey-market peptides safe? What the lab data shows (and what it can't)

Somebody asks "is it safe" and means two different questions at once, usually without noticing. One is: does this vial contain what it says it contains? The other is: is it safe for a human body? A lab report can answer the first question, sometimes well. It cannot answer the second one, ever, and no amount of testing changes that. This piece is about drawing that line clearly, using our own data as the example, without pretending either side of it doesn't matter.

That's not a call a lab report is equipped to make. What follows is what the numbers actually say, and — just as important — what they were never designed to say.

What lab testing can measure

A certificate of analysis, run by an independent lab, can typically speak to a specific handful of things about the sample tested: whether the molecule present matches the one claimed (identity), how much of what's in the vial is that molecule versus something else (purity), how much of the compound is actually there compared with the label (dose accuracy), whether bacterial contamination markers are present (endotoxin), and whether the sample is free of live microbial growth (sterility). Each of those is a real, useful measurement. None of them is "safety" as a whole.

Here's what our database — 1,527 vials from 67 vendors, tested by 20 independent labs (mostly Janoshik, 1,065 results, followed by Peptide Test/TrustPoint, Freedom Diagnostics, Chromate, and others), current as of September 28, 2026 — actually shows on each of these fronts.

Purity. Today, of 1,478 vials with a purity result, 1,341 — 91% — came back at 99% or higher, and 14 tested under 95%. On this specific measure, most tested vials in our data are clean: whatever is in the vial is mostly the labeled compound, not filler or degradation byproduct. That's a real, if narrow, piece of good news, and it's worth saying plainly rather than burying it.

Dose accuracy. This is where the picture gets more complicated, and where purity alone can mislead you. Of 1,441 vials with a measured amount against a labeled amount, the median result ran about 6.5% over label — vendors skew slightly toward overfilling rather than underfilling, on average. But averages hide a real spread:

About 5% of tested vials (75 of 1,441) came in more than 10% under their labeled amount, and a smaller but real slice — 52 vials — combined 99%+ purity with more than 10% underdosing, 20 of those by more than 20%. That combination is the trap: a vial can look immaculate on the purity line and still shortchange the buyer badly on quantity. We cover which specific compounds run into this most often in a separate piece on underdosing, because the pattern is not evenly distributed — GHK-Cu and BPC-157 skewed toward significant overfill in our data, for instance, while a smaller number of vials across several compounds came in badly under label.

Endotoxin. This one measures a marker for bacterial contamination, and it's the test that most clearly is not routine. Of 202 peptide results in our data with an endotoxin figure (roughly 28 vendors, January 2025 through September 2026, dominated by tirzepatide and retatrutide samples), about 12% — roughly 24 results — came in at 100 EU or higher. The highest single reading was 26,410 EU per vial for an epitalon sample, and on a single day, eight of eleven NAD+ vials from one vendor, each from a different batch, came in above 500 EU per vial. We're not going to translate that into a health claim or compare it to a pharmacopeial injectable limit dressed up as a "safe" threshold — that's a different kind of claim than a lab report supports, and units aren't even consistent across these results (some are per vial, some just say "EU" with no stated basis, which makes cross-vial comparison unreliable). What we will say: these are the vials buyers specifically chose to pay extra to test for endotoxin, which is very likely not a random sample of the market — people tend to test when something already seems off, or when a vendor's reputation is already shaky. Treat this figure as a signal that the risk is real and unevenly distributed, not as an average condition of every vial sold. We go deeper into the endotoxin numbers, unit inconsistencies, and which compounds show up most in that dataset in a dedicated piece.

Sterility. The smallest and most alarming-looking dataset: 17 samples tested for total aerobic microbial count and total yeast/mold count, and 14 of them failed at least one measure, 10 failed both. All 13 tirzepatide and retatrutide samples in that group failed at least one count. We're stating that number because it's real and it's ours, but the caveat matters as much as the figure: 17 samples is a tiny slice of the market, almost certainly skewed toward vials someone already suspected, and a failed microbial count is a lab finding about that sample — not a clinical diagnosis about what happens if it's used.

The smaller tests, briefly. A few other measurements show up in our data less often but are worth naming. TFA (residual trifluoroacetic acid, a byproduct of some purification methods) turned up in 26 results, nearly all reading "0" or "not detected," with two tirzepatide results reporting a small amount (1.4 mcg). Seven heavy-metal screens all came back "not detected." Ten LC-MS impurity screens typically flagged mannitol alongside the target peptide and minor degradation products — mannitol is a common bulking agent in freeze-dried vials and isn't inherently a red flag on its own. None of these are large enough samples to generalize from, but they round out the picture: where testing exists at all, gross contamination by these specific measures is rare in what buyers have chosen to test.

What lab testing cannot tell you, at all

This is the part that gets skipped in most conversations about "is it safe," and it's the part that actually determines whether using an unapproved compound is a good idea for a given person.

A lab report cannot tell you whether the compound itself is safe for human use at any dose. Identity and purity testing confirms what molecule is present; it says nothing about that molecule's pharmacology, its interaction with a person's other medications or conditions, or what happens with repeated use over months or years. And where a compound has been through clinical trials, those trials studied a specific pharmaceutical product, made and handled under controlled conditions. A research-labeled vial isn't that product, however clean it tests.

It cannot tell you anything about the vial you specifically hold. A report documents one sample from one batch, tested on one date. It's evidence about that batch, useful as a signal about a vendor's track record over many reports, but it is not a guarantee extending to the vial in your hand, which could be from a different production run entirely — a point we cover in more depth in our guide to reading a COA.

It cannot account for how a compound is used. Dosing, reconstitution, storage, and injection practice all affect outcomes in ways no lab measurement captures, and that's deliberately outside what this article — or this site — covers.

It cannot tell you about long-term effects. Nobody has run the kind of extended human trials on most of these compounds, at these routes of use, that would generate that data. The FDA hasn't approved them for these uses precisely because that evidence base doesn't exist yet for most of them.

And a clean report from a reputable lab does not retroactively make an unapproved drug an approved one. Put simply, independent testing reduces one category of risk — getting a mislabeled or contaminated product — without touching the underlying question of whether using an unapproved compound is medically appropriate for someone at all. Those are separate risks, and no COA closes the gap between them.

The picture varies a lot by compound

"Grey-market peptides" isn't one product, and the safety-relevant testing data reflects that. As of September 28, 2026, tirzepatide and retatrutide dominate our dataset (788 and 336 vials respectively, out of 1,527 total) and skew toward the more reassuring end of the numbers above — a median of +5.4% versus label for tirzepatide, +8.2% for retatrutide, with failure counts (empty, under 95% pure, more than 20% under or 50% over label) in the low single-digit percentages for both. Smaller categories look rougher in our data: GHK-Cu, with 29 tested vials, had 10 come in more than 50% over label, counted as a fail at that level in our scoring; BPC-157 ran a median of nearly 18% over label across a much smaller sample of 41 vials. None of that is a claim that any specific compound is inherently better or worse made — sample sizes for the less-tested compounds are small enough that a handful of bad vials moves the numbers a lot. It's a reason not to generalize from "grey-market peptides" as a single bucket to a specific compound you're researching.

Why "Reddit says it's safe" isn't evidence, either

A specific version of this question shows up constantly in searches and forum threads: is it safe, according to the people who've already used it? Anecdotal reports from other buyers are a different category of information from lab data, and worth naming as such rather than dismissing outright. A thread full of people saying they used a compound without an obvious adverse event is real information about what didn't visibly go wrong for those specific people, at whatever dose and frequency they used, over whatever time period they've been paying attention. It is not a substitute for controlled data, it doesn't capture effects that take a long time to show up, and it says nothing about a person with a different health profile, different medications, or a vial from a different batch. Community sentiment can be a useful supplement to a vendor's testing history — a burst of complaints is often the first sign something changed with a batch, well before any lab confirms it — but it's a signal about reputation and reporting, not a safety verdict.

The honest answer

So: are grey-market peptides safe? Based on what independent testing actually covers, the honest answer is that testing can tell you a compound is probably what it claims to be, roughly the dose it claims to be, and free of two specific contamination markers — sometimes. It cannot tell you the product is safe to use, because "safe to use" is a medical question that identity and purity testing was never built to answer, and because even a strong batch-level track record doesn't guarantee the specific vial someone is holding.

What we can do is be precise about what a passing lab report actually proves, and what it leaves completely untouched. If you're trying to reduce the risk that comes specifically from an inaccurate or contaminated vial — as opposed to the separate risk of using an unapproved compound at all — a vendor's testing history, read carefully and skeptically, is the best tool this market currently offers.

Here's how the ranked vendors look right now:

A rank isn't a safety guarantee — check the vendor's page for results on the specific compound you're looking at, not just the overall score, and don't skip a look at how recent the reports are. If you've already got a vial and it comes back bad on your own test, that's useful information too: stop using that batch, flag it to whatever group or forum you bought through, and submit the report so the next buyer sees it.

Database figures without a date update automatically from VendorVerify's lab results; dated figures are snapshots. This article is not medical advice.