The Vendor COA Is an Advertisement

A certificate of analysis, or COA, is the document a research-peptide vendor posts to show that what is in the vial matches the label, and in the gray market it functions the way an advertisement does: it presents the product at its best, on the vendor’s terms, in a format designed to close the sale. This article teaches COA literacy, starting with the four documented ways these documents mislead, then the QR code and paper-trail checks, the lab tests that carry the weight, and the group-testing economics that make verification a $20 to $40 habit. Everything below is community-sourced from the peptide research community’s testing discourse, so quotes are labeled as anecdotal where they reflect community or personal experience, and none of it is medical advice; these compounds are for research use only.

I · Anatomy of a certificate of analysis

The stakes explain why the community reads these documents so carefully, because a COA is often the only evidence between a buyer and an injectable vial. The compounds at the center of this discourse are potent enough that Jastreboff’s phase 2 obesity trial, published in the New England Journal of Medicine, reported a 24.2% mean weight loss at 48 weeks on the 12 mg dose of retatrutide, which is the kind of outcome that makes people willing to inject unregulated vials and the reason verification matters when they do. ² A COA is the artifact that supposedly closes that gap, since it reports the mass and purity of a tested sample alongside the product name, the batch number, the testing lab’s identity, a signature line, and increasingly a QR code that links back to the lab’s records. ¹

The community’s most experienced testing advocate, a knowledgeable amateur who has spent years coordinating group tests and reading lab reports, frames the genre in a line that circulates constantly: [anecdotal]

I like to think of vendors COAs as an advertisement. “This is what I want to sell you.”

Community discourse, peptide sourcing corpus, 2025

That framing is the key to COA literacy, because an advertisement describes the product the seller wants you to believe in, and the certificate does the same thing with laboratory formatting. The same educator’s shorthand rule is to “treat a vendor test as a commitment rather than a truth,” and a separate sourcing-cluster commentator states the corollary plainly: “Because the COA is a marketing tool. It convinces people like you that this is a good product to buy.” [anecdotal] ^{1 3}

The document is worth reading even after you accept that framing, because the fields tell you what the test covered. The product name tells you what the vendor asked the lab to check, the batch number tells you which production run the sample came from, the mass result tells you how much peptide was in the vial, and the purity result tells you how much of that mass is the target molecule rather than impurities, while the lab name and QR code tell you who to call if you want the raw data. ¹

The reading habit that follows is to note the batch number and the lab identity first, since those are the two fields the verification stack uses. ¹

II · Four ways a certificate of analysis lies

The opening observation in the testing threads is blunt, since one of the most experienced sourcing commentators states that “There basically isn’t a vendor who hasn’t sent out bad products, often despite providing good and real COAs.” [anecdotal] ⁶ That sentence is the reason the community tests at all, since a good and real COA can accompany a bad product through any of the four failure modes cataloged in the sourcing clusters. ^{4 5 7 8}

The first way is outright fabrication, which means the document itself is fake or photoshopped, and in the extreme case it was never issued by the lab named on the page. The fakery catalog includes “Fake Jano COAs on their site. QR code links to a COA for a completely different vendor,” plus “fake GMP paperwork along with a photoshopped Jano test,” and one QR-code check that exposed the edit: “If you click the QR code you can see that the COA he posted is a photoshop.” [anecdotal] ⁴ A fabricated COA is the easiest to catch with a QR scan and the hardest to argue with when the buyer does not scan it.

The second way is batch switching, where the document is real but describes a different production run than the one you receive. The canonical explanation is that “the vendor showed a COA for one batch and sent the customer a kit from a different batch,” and the community’s testing advocate makes the same point from the buyer’s angle: “the concern is that your favorite vendor may be sending you a different product than the one they sent to Jano.” [anecdotal] ⁵ The most quoted version comes from the same source: “Janoshik is very legit, but it’s easy for the vendors to cheat a test. Who’s to say that the product they send to you is the same as the one they sent to Janoshik?” [anecdotal] ⁵ Batch switching is why the batch number on the COA must match the batch number printed on the kit, and why a COA without a batch number is close to useless.

The third way is warehouse mislabeling, where the vendor’s COA is genuine and accurately represents the product they intended to sell, and the failure happens downstream. The sourcing-cluster explainer puts it directly: “the vendor will often have a real COA of their own that accurately represents the product they intended to sell you… the problem being that their warehouse people screwed up and mislabeled stuff so you got sent the wrong thing.” [anecdotal] ⁷ Documented examples include the incident where “even though the COAs were legit everybody got the wrong (mislabeled) thing,” plus a kit that was “somebody else’s selank labeled as reta” and the widely shared “6mg” reta vials that tested as 4.5 mg of tirzepatide with “COAs from like half a dozen labs.” [anecdotal] ⁷

The fourth way is the re-branded reprint, where a reseller buys the upstream vendor’s test and prints their own name on the header. The community’s most mechanistic commentator describes the economics with dry precision: “half of these guys are actually testing their products and aren’t just getting a copy of their vendor’s COA with their own name on the header (a service that the testing labs provide for a small fee).” [anecdotal] ⁸ The fee is documented at $35, since one thread records the exchange “You mean pay Janoshik $35 to put your name on somebody else’s test?” and another sourcing packet notes the labs “charge a small amount for this service.” [anecdotal] ⁸ The reprint practice is why a COA with the reseller’s logo on it proves little about who tested the product, and why one commentator’s reply covers the whole genre: “I can print a sticker that says my vials are from that batch too. Doesn’t make it so.” [anecdotal] ⁸

All four ways collapse into a single question that the reader can ask of any document: does this certificate describe the vial you are holding? If the batch number matches, the lab is real, the QR resolves, and the product name lines up, the document has earned its place as a starting point. ⁸

III · The QR code and the paper trail

The QR code turns verification into a thirty-second task, because scanning it should land on the testing lab’s own page for that document, and the community’s instructions are exact: check that the QR code links to the lab’s site, and treat a code that resolves anywhere else as a red flag. ¹⁰ The failure modes are documented, since QR codes have been found linking to “a COA for a completely different vendor,” and the photoshop case above was caught because the scanned code exposed the edit. [anecdotal] ^{4 10} The inverse tell is equally strong, since a QR code cropped or scrubbed from the document is itself a warning sign. ¹⁰

The rest of the paper trail gets the same treatment, starting with the signature line, since a COA with no named signer is a document nobody will vouch for: “There’s not even an actual person’s name on the signature line.” [anecdotal] ¹¹ The community also warns that “I’ve seen people alter vendor names or results,” which is why the QR verification matters more than the printed header, and why the lab identity deserves a search of its own. [anecdotal] ¹⁰ Some of the entities issuing certificates “They’re not actually a lab. That’s just a business name,” and the testing-discourse commentators point out that “everybody with access to Adobe Acrobat and a telephone” can produce a plausible-looking PDF, so the lab’s track record matters as much as the document’s formatting. [anecdotal] ¹¹

The deeper lab check is reproducibility, since a real testing operation produces results other labs can replicate, and community trust rests on that. One lab in the corpus is described as having “regularly put out test results that are not reproducible by any other labs,” with a documented case where a 91.4% purity COA from that lab retested clean at reputable labs twice. [anecdotal] ¹² The community’s verdict on the bad actors in the lab space is delivered with characteristic dryness: “Some of them aren’t testing products correctly. That’s the fun part.” [anecdotal] ¹²

The paper-trail habit is cheap and repeatable, so the community treats it as the entry-level check: scan the code, confirm the lab, confirm the batch, and only then decide whether the document is worth building on. ¹⁰

IV · Mass and purity, then endotoxin

The community’s testing doctrine ranks the lab tests by value with unusual agreement, since the most experienced commentators converge on the same hierarchy, with mass and purity at the top and sterility and heavy metals near the bottom while endotoxin sits between them. ¹³ The doctrine statement that gets quoted most is the sourcing-cluster rule that “The standard mass and purity test is the only test that really has much value,” and the same packet’s blunt summary of the tier list reads: “Mass and purity testing is important. Endotoxin testing is a good idea. Heavy metals testing is a waste of your money.” [anecdotal] ^{13 17}

Mass testing is the fill check, since it measures how much peptide made it into the vial, and the community’s consistent finding is that fills run slightly over rather than under: “they’re usually slightly high,” with a “baker’s dozen” overfill of 5% to 10% treated as the norm and purity targets of “98-99% minimum” as the quality bar. [anecdotal] ¹³ The overfill lore includes a GHK-Cu batch where “50mg” vials tested at 100 mg, while COAs over 100% purity get a “caveat emptor” shrug and a 99.88% purity number without batch context earns the warning “Careful with that COA.” [anecdotal] ¹³

Endotoxin is the test whose status changed during the corpus, because endotoxins are fragments of dead bacterial cell walls that survive filtration, and for years the community never saw one fail: “I’ve never seen an endotoxin test come back high.” [anecdotal] ¹⁴ The FDA threshold for a parenteral drug is 5 EU per kilogram of body weight per dose, which the community works out to 29.167 EU per mg for retatrutide at 12 mg for a 70 kg adult, about 350 EU per dose, and most posted “failures” turn out to be 3% or less of that limit. [anecdotal] ¹⁴ Then the story shifted, since multiple vials from a 30 mg tirzepatide batch failed at about three times the FDA limit, and one of the most experienced testing advocates marked the change in the sourcing packets: “Until very recently we had never actually seen a Chinese peptide fail an endotoxin test, but multiple vials from a 30mg tirz batch just failed endotoxin testing so that seems like a more worthwhile test than it did just a few weeks ago.” [anecdotal] ¹⁵ His one-line summary of the new reality is “Endotoxins are unfortunately an issue now.” [anecdotal] ¹⁵

Sterility testing sits at the bottom of the value list for a structural reason, since a sterility failure is often vial-by-vial rather than batch-wide, which means a pass on one vial says little about the next. The community calls it “a completely useless test because it’s often a vial-by-vial rather than a batch issue,” and the replacement is mechanical rather than analytical: “It’s cheaper to filter than to pay for sterility tests, and more effective.” [anecdotal] ¹⁶ The standing practice is to run each reconstitution through a 0.22 micron sterile filter into a fresh sterile vial, since “You’re better off treating every single vial as a possible failure and just filtering. They’re all passes after you filter.” [anecdotal] ¹⁶ Heavy metals are deprioritized for a different reason, because they “would pretty much require malicious intent” in a production process that has no use for them, and the community’s dismissive line is that “You’d get a lot more bang for your buck worrying about the heavy metal content in your protein shakes.” [anecdotal] ¹⁷ The one expensive test worth knowing about is LCMS identification for a mystery vial at about $600, a price point one commentator meets with “Might be more economical to chuck it.” [anecdotal] ¹⁸

The ordering is the takeaway, since a limited testing budget goes to mass and purity first, with endotoxin added after the 2026 failures and sterility solved by filtration rather than by a test. ^{13 16}

V · Group testing and the four-vial rule

The economics of verification are the reason the community tests at all, because a solo mass and purity test runs about $300 for the first vial plus $60 for each additional vial, near $350 all-in with shipping, while a group test costs $20 to $40 per participant. ¹⁹ The standard mechanics come from the most-cited testing thread: “The most common approach to testing is group testing. Finding 5-40 other people who bought the same batch, send 1-4 vials in for testing, split the cost.” [anecdotal] ¹⁹ The worked examples in the corpus include a 30-member group sending three vials for a total of $480, about $20 per member, and one experienced tester’s group test of R50 kits that matched the vendor COA at $18 per participant. [anecdotal] ¹⁹

The four-vial rule is the second half of the doctrine, since a single vial can be the one fill-nozzle miscalibration in an otherwise clean batch, and the sourcing cluster states the instruction directly: “Test 4 vials from a batch, not 1 from a kit,” because that is what catches miscalibrated fill nozzles and batch variance. [anecdotal] ²⁰ The observation that made the rule famous is the finding that “multi-vial tests occasionally bifurcate into two different clusters,” which happens when a vendor ships two batches in one order, and the documented extremes include a US finishing operation with 40% fill variation within a single batch and a finisher whose fills varied about 2.5% off the mean against another at 0.01%. [anecdotal] ²⁰ One tester’s own three-vial run on vials labeled 50 mg returned 53.06, 52.92, and 52.92 mg, a 6% average overfill with infinitesimal vial-to-vial variance, which is the shape of a healthy batch. [anecdotal] ²⁰

The protocol details matter as much as the numbers, since a mass measurement requires the whole vial: “Send a whole vial or you don’t get the mass measurement,” and partial-sample testing defeats the purpose. [anecdotal] ²¹ The community’s gold-standard lab runs blind, since the operator “doesn’t know the spec for the peptide or the mass” when the sample arrives, and the corpus records known-bad kits being sent in to validate that a lab catches problems. [anecdotal] ²¹ Vendor warranties often hinge on the lab used, since many vendors honor guarantees only with results from the community-standard lab, a quiet admission that vendor-controlled testing is the weaker link. ¹⁹

The verification loop closes with the batch match, since the community’s rule is that “batch-matching COAs for your kit is the only way to trust,” which means the tested batch number has to be the batch number printed on the kit you hold. [anecdotal] ²² A COA for a batch you do not have is a picture of somebody else’s product, and the four-vial group test exists to close that gap for the batch you bought. ²²

VI · The verification habit

Line up the full stack and the pattern is a habit, since each check is cheap and each one narrows the distance between the document and the vial. The COA is the starting point, “a good starting point” in the community’s own words, and from there the reader confirms the document’s origin with the QR code, ties it to a specific run with the batch number, and measures its claims against what was received with a group test. [anecdotal] ^{3 22} The final layer is mechanical, since the 0.22 micron filter at reconstitution handles the sterility question that no certificate can answer for your individual vial. ¹⁶

The habit extends past certificates, because the same community that fact-checks vendor COAs also fact-checks its influencers, and the corpus records it tracing one prominent influencer, Trevor Bachmeyer, whose “reta requires carbs” claims were built on fabricated ChatGPT citations that collapsed under the same scrutiny. [anecdotal] ²³ That episode is the verification habit in miniature, since a document asserting a fact is still a claim until it checks out, whether the claim sits in a certificate of analysis or an influencer’s screenshot of one.

The open questions are honest ones, because the verification stack is young and the market keeps moving. The 2026 endotoxin failures raise the question of whether more batches will fail as testing normalizes, while the per-vial sterility risk stays structurally untestable at any price and the reprint economy keeps the community-standard labs as the most reliable arbiters until more of them publish searchable verification pages. ²⁴ The one conclusion the threads support without qualification is the doctrine this article started with: the certificate describes the product the vendor wanted to sell, and verification is how you learn what the vendor shipped. ²⁴