A Certificate of Analysis is the single document that stands between a buyer and a guess. It is also one of the most misread documents in the peptide market — not because the numbers are wrong, but because they answer narrower questions than most people assume. A COA can be entirely accurate and still leave a buyer with the wrong impression, simply because they read the top line and stopped.
To make this concrete, we will walk through a real report, shown in full to the right: a third-party Certificate of Analysis issued by Kovera Labs for a single lot (NX20100528261). It happens to be an in-house Peptinexia lot — we are using it because we can publish it in full, not as an endorsement of the product. The reading skills transfer to any COA you are handed.
Identity: is this the right molecule?
Identity testing confirms that the material in the vial is the peptide named on the label — no more, no less. It is typically established by mass spectrometry, which measures the molecular weight of the compound and compares it to the theoretical mass of the intended sequence. On our example report, identity is confirmed by LC-MS against the stated molecular formula.
Identity is a yes/no question, and it is the most important one. A high purity number is meaningless if the underlying molecule is not what you think it is. When identity is absent from a report entirely, no other number on the page can be trusted.
Purity: how much of the material is the target compound?
Purity is almost always measured by high-performance liquid chromatography (HPLC), which separates a sample into its component peaks and reports the target peak as a percentage of the total. Our example reports a batch-average purity of 99.746% by RP-HPLC (C18 column, DAD detection at 214 nm), with the chromatogram printed on the certificate — a single dominant peak, which is what “high purity” looks like on paper.
Two cautions. First, purity is method-dependent: the wavelength, column, and gradient all affect what gets counted, so a purity figure without a stated method is a number without a unit. Our example states all three. Second — and this is the misread that costs buyers the most — purity is a ratio, not a quantity. It tells you what fraction of the material present is the target. It tells you nothing about how much material is actually in the vial.
Net content: how much is actually there?
Net content — sometimes called peptide content, fill, or mass — answers the question purity cannot: how many milligrams of peptide are in the vial versus what the label claims. On our example, the label reads 10 mg and the measured batch-average net content is 12.63 mg. This lot is over-filled — the buyer receives more than the label promises.
A purity percentage and a fill weight are different facts. This lot is 99.7% pure and holds 12.6 mg where the label says 10 — read only the first and you would never know the second.
That is the point worth sitting with. Purity (99.7%) and net content (12.63 mg) are independent numbers, and neither predicts the other. Here the fill is generous. Elsewhere in the market it is not — underfill, where a vial contains meaningfully less than its label, is a common and under-discussed discrepancy, and it hides comfortably behind an impressive purity figure. Whichever direction it runs, the lesson is the same: if a COA reports purity but omits net content, the one number that tells you the quantity you are receiving has been left off the page.
Heavy metals, endotoxin, and sterility
Beyond identity and quantity, a thorough COA reports on contamination. Our example lot carries separate analytical reports for each:
Heavy metals, by ICP-MS against EPA-referenced methods, with lead, arsenic, cadmium, and mercury each reported below defined acceptance limits, and spike-recovery values shown to confirm the method worked on this matrix. Bacterial endotoxin, by kinetic LAL assay under USP <85>, reported at less than 0.20 EU/mL against a 0.5 EU/mL limit. And a 14-day sterility screen reporting no microbial growth. Each of these is meaningful only because the report states the method and the limit, not merely “pass.”
The stronger a testing panel, the more of these you will see broken out into their own documents rather than compressed into a single “pass.”
Provenance: does the report trace to the lot?
The most important structural feature of a COA is not any single number — it is whether the document ties to a specific lot. Our example does this in three ways worth imitating as a checklist. It carries a unique lot and report identifier (NX20100528261 / KVR-2026-B136E2). It reports batch conformity vial-by-vial — Vial 1 at 99.679% and 12.44 mg, Vial 2 at 99.813% and 12.83 mg — rather than a single abstract figure, so you can see the spread. And it prints a verification route (a QR code and an access code) that lets a holder confirm the report against the issuing lab directly.
A generic COA — one that describes “the product” in the abstract, with no lot identifier, no test date, and no issuing laboratory — describes an idea, not an object. The presence of a lot number, a test date, per-vial results, and a named third-party lab tells you more than any purity figure.
A reading checklist
When a COA lands in front of you, read it in this order:
- Is there a lot number, a test date, and a named laboratory?
- Is identity confirmed, by a stated method?
- Is purity reported with its method?
- Is net content reported — and how does it compare to the label, in either direction?
- Are contamination tests reported with methods and limits, not just “pass”?
- Can you verify the report against the issuing lab?
A report that answers all six is doing what a COA is for. A report that answers only the flattering ones is doing something else.
