A certificate of analysis is the single most informative document a peptide vendor publishes, and almost nobody opens one. People look for the badge on the product page that says the thing was tested, see it, and stop there.
Part of the reason is that the document looks impenetrable. It is a wall of abbreviations, a chart nobody explains, and a column of numbers with no obvious meaning. So here is the whole thing in plain language: what each section is, what the tests actually measure, and the order to check things in. No chemistry background assumed.
By the end you should be able to open a certificate, spend ten minutes on it, and come away knowing something real about what you are looking at.
What a COA is, and what it is not
A certificate of analysis is a report from a laboratory describing what it found when it analyzed a sample. That is all it is. It is a record of measurements on one sample of one batch at one point in time.
Four things follow from that definition, and each one matters.
A COA describes a batch, not a product. A certificate for lot BPC10-08 says nothing about lot BPC10-09. Batches are made separately and can turn out differently. If the document you are shown is not for the lot number printed on the vial you receive, it is background reading, not verification.
A COA is only as independent as the lab that issued it. A report produced in-house by the seller is a self-assessment. That is not automatically worthless, but it is a different kind of evidence from an outside lab with its own reputation at stake.
A COA only covers the tests it actually ran. This sounds obvious. It is the single most common gap I find. A document can be entirely genuine, issued by a real lab, and still not contain the test the product page is advertising.
A COA is a claim until you verify it. A PDF hosted by a vendor is a file the vendor controls. The question that matters is whether the issuing lab's own records say the same thing.
Finding the document in the first place
Before you can read one you have to get one, and how hard that is tells you something on its own.
The good version is a link on the individual product page, or a certificates library you can search by product and lot. The mediocre version is a single page of PDFs with no lot numbers, so you cannot tell which batch anything belongs to. The bad version is a badge image that says "third-party tested" and links nowhere at all.
A badge is a graphic someone added to a web page. It is not evidence of anything. The document is the evidence. If you cannot find a document, that is your answer, and you can stop there.
The anatomy of a certificate
Nearly every real COA has the same four blocks, usually in this order down the page.
The header: who, what, and when. The lab's name and address, the client the report was prepared for, the product name, the lot or batch number, and a set of dates. There should be three dates, not one: when the sample was received, when it was tested, and when the report was issued. One lonely date is worth noticing, because it collapses a timeline into a single unverifiable point.
The client line is the one people skip, and it matters more than it looks. It names who paid for the test. If it names a company other than the vendor you are buying from, that certificate documents somebody else's material.
The methods. Usually a column beside each result, naming the analytical technique used. This is what turns a number into a measurement. A purity figure with no method beside it is a number with no provenance.
The results table. Typically three columns: the test, the specification, and the result. The specification column is the one that does the real work, and it gets its own section below because it is the most commonly missed idea on the whole page.
The signature and verification block. The name and credentials of whoever signed off, and ideally a QR code or access code that resolves to the lab's own copy of the same certificate. That last part is the difference between a document you are shown and a document you can confirm.
If a certificate is missing one of these blocks entirely, that is a finding in itself.
The four things a certificate measures
This is the part nobody explains, so here it is properly.
Identity: is this the right molecule?
Usually confirmed by mass spectrometry, abbreviated MS. The instrument measures the mass of the molecule. Every peptide has a known expected mass based on its amino acid sequence, so if the measured mass matches the expected mass, you have good evidence the molecule is what the label says.
What you want to see is the actual trace or spectrum reproduced on the certificate, not just the word "conforms." A conclusion with no data behind it is an assertion.
Identity is a separate question from purity. A sample can be unambiguously the right molecule and still be mostly something else by weight. Which brings us to the next two.
Purity: how much of the peptide material is the right peptide?
Usually measured by HPLC, high-performance liquid chromatography. The sample is pushed through a column, the different components in it travel at different speeds, and they come out the far end at different times. A detector records what emerges over time, and the output is a chromatogram: a flat baseline with peaks rising out of it. Each peak is a component.
Purity is normally calculated as area percent. The software measures the area under the main peak, divides it by the total area of all the peaks, and reports that as a percentage. "99.1% by HPLC" means the main peak accounts for 99.1 percent of the total detected peak area.
A good chromatogram looks like a flat line with one tall, narrow, symmetrical peak and very little else. Small bumps elsewhere are related impurities, usually truncated or modified versions of the peptide left over from synthesis.
Two things worth knowing, because they are where this number gets oversold:
- It only counts what the detector can see. Most peptide HPLC uses ultraviolet detection, commonly at 214 nanometers, which responds to the peptide bond itself. Anything that does not absorb at that wavelength is effectively invisible to the measurement. This is why the detection wavelength belongs on the certificate.
- Salts and water are invisible to it. They do not produce peaks. So a vial can legitimately be 99 percent pure by HPLC and still be substantially less than 99 percent peptide by weight.
That second point is the single most misunderstood thing in this whole category, and it has its own test.
Net peptide content: how much of what is in the vial is actually peptide?
Purity answers "of the peptide material present, what fraction is the right peptide." Net peptide content answers a different and more practical question: "of the total mass in this vial, what fraction is peptide at all?"
The rest is not contamination in a sinister sense. Synthetic peptides are usually purified in a way that leaves them paired with counterions, most often trifluoroacetate from the purification step, plus some residual water the powder picks up. Both have mass. Neither is peptide.
So a vial can be 99 percent pure and 78 percent peptide by mass at the same time, and both numbers can be completely honest. They are answering different questions. Net peptide content is typically determined by amino acid analysis or by nitrogen determination.
Most certificates in this market do not report it. When one does, that is a meaningful sign of a more serious testing package, and it is worth checking what the number actually is rather than just noting that it exists.
The safety panel: what else is in there?
These are separate tests and each one is either present or absent. Absence is information.
- Sterility asks whether viable microorganisms grow from the sample. The relevant reference is USP General Chapter 71.
- Bacterial endotoxin, usually by a method called LAL, looks for fragments of bacterial cell walls. These are worth testing separately because they survive processes that kill the bacteria themselves. USP General Chapter 85 covers it. A serious report states the sensitivity of the assay and shows the replicates.
- Heavy metals, usually by ICP-MS, which ionizes the sample in a plasma and identifies elements by mass. USP General Chapter 233 is the procedures reference. Look for actual quantified values against limits, not a bare "complies."
- Residual solvents, the solvents left over from synthesis and purification, usually measured by gas chromatography and referenced to the ICH Q3C guideline.
A certificate carrying none of these is not necessarily a bad certificate. It is a narrow one, and you should know that is what you are holding.
The specification column, and why PASS can be meaningless
This is the sharpest single idea on the page, and it takes one minute to learn.
Every row in a results table should have three parts: the test, the specification (also called the acceptance criterion, the limit that had to be met), and the measured result. PASS means the result fell inside the specification.
So PASS only carries information if a specification exists. Some certificates mark a row's specification as "Report Only." That means exactly what it sounds like: measure it and write it down, with no limit attached. A "Report Only" row cannot fail. The result can be anything at all and the row will still read PASS, and the certificate's overall verdict will still read PASS.
I have seen a batch where three separate vials were sampled and the measured content came back at roughly 1531, 1557 and 104 against a 1500 label. One vial held a small fraction of what the other two did. Every row said PASS, because that test was specified as "Report Only."
Nothing failed, because nothing could fail. The vendor publishing the raw number was doing the honest thing; the specification is what made the verdict meaningless. So whenever you see PASS, look left. If there is no limit in the specification column, the word PASS is decoration.
A ten-minute check, in order
Cheapest checks first, so you can stop early when something obvious is wrong.
- Does a certificate exist for this specific product? Not for the vendor generally. For the item you intend to buy. Coverage across a catalog varies enormously.
- Does it match your lot? Compare the lot number on the document against the number printed on the vial. Before ordering, you are really checking whether the vendor publishes per-lot documents at all.
- Who is the client? The certificate should be made out to the vendor you are buying from.
- Who is the lab? A named lab you can look up, not "an independent third-party laboratory" with no name attached. Then go look it up.
- If accreditation is claimed, does it check out? ISO/IEC 17025 is the standard for testing laboratories, and accreditation under it is granted by a named accreditation body, never self-declared. Find the accreditation bodies through the ILAC signatory search; most publish a searchable directory of the labs they have accredited. Two minutes, and worth it, because accreditation is the kind of claim that sounds like a credential and often is not one.
- Is there a purity result, and by what method? You want HPLC with the chromatogram shown and the wavelength stated, plus identity by mass spectrometry with its trace.
- Does each result have a specification? Check the column. Remember "Report Only."
- Can you verify it at the lab? Take the access code or QR link to the lab's own portal, and compare every field against the vendor's copy rather than just confirming the document exists.
- What is not on the panel? Sterility, endotoxin, heavy metals, residual solvents, net peptide content. Note what is missing.
Red flags, in rough order of how often they turn up
Every one of these is drawn from certificates I have actually read while reviewing vendors. Most are not fraud. Most are a marketing department outrunning a lab report.
- The advertised test is not in the document. Product page promises HPLC purity, the certificate tests something else entirely. The most common problem by a wide margin, and it survives because the documents are real, the lab is real, and almost nobody opens them.
- A purity badge with no purity test behind it. Related, but worth separating: the number in the badge appears nowhere in the paperwork it links to.
- A credential that evaporates when checked. "ISO-accredited laboratories," plural, where the reports all come from one lab and no accreditation appears anywhere on that lab's own site.
- The client line names someone else. Often an upstream supplier, with the vendor republishing its document. The testing can be perfectly real and still not be evidence about the stock you would receive.
- The vendor's copy and the lab's copy disagree. Rarer, because it takes effort to find, and the reason step 8 exists. I have seen a vendor PDF report every heavy metal as "Not Detected" while the lab's own record for the same sample carried quantified values, one of them at 88 percent of its limit. Everything passed. But "Not Detected" and a measured number are different statements, and the difference ran in the direction that flattered the seller.
- A single date instead of three.
- PASS with an empty specification column.
Worth adding, because it keeps the previous point honest: I have also found a lab record with an extra test row that the vendor's published copy left out, meaning the vendor was understating testing it had paid for. A discrepancy is a question, not automatically a verdict.
What a certificate still cannot tell you
Worth being honest about the ceiling.
A certificate covers the sample that reached the lab. It does not prove that the sample came from the batch your vial came from, that storage and shipping preserved the material, or that what is in your particular vial matches what was tested. Those gaps close only by matching the lot number yourself when the package arrives, and even then you are trusting the chain between the lab bench and your doorstep.
It also tells you nothing whatsoever about whether the compound is worth anything. A flawless certificate on a peptide with thin evidence behind it is a well-documented vial of something nobody has studied properly. Those are two separate questions and they need two separate investigations.
That is why how I review vendors separates a first-look profile from a hands-on one, and why the Papa Score rubric scores a verified batch match separately from published documentation.
Where to go next
New to the category? The companion piece is what a peptide actually is, including how this market reached its current state and why documentation is the only real leverage a buyer has.
Researching a specific compound? That is the other half of the job, and a certificate cannot help you with it. The evidence behind each compound is its own question, with very different answers for BPC-157, semaglutide and ipamorelin.
Want this standard applied seller by seller? That is what the vendor comparisons are for. Each review states which certificates I opened, which lab issued them, what I could verify, and what I could not.
And the general rule, which has held up across every vendor I have looked at: the badge is marketing, the certificate is evidence, and the lab's own copy of the certificate is the only version nobody selling to you controls.