The short answer
A certificate of analysis, or COA, is a laboratory’s record of tests on a sample tied to a stated lot. It tells you what someone measured, by what method, and what number they wrote down.
It does not tell you that a substance is safe, that it works, or that it is suitable for use in people. It is not a clinical trial or a government approval.
What a COA is actually claiming
Most COAs are organized around a few questions. What is this material? How much of the named substance appears to be present? What else showed up? Which methods were used? Who tested it, and when?
In medicines made for people, those questions sit inside a larger system of manufacturing controls.
A final test is not the whole guarantee.
A COA handed to you in some other context can borrow the same words, such as identity, purity, and appearance, without inheriting the system around them.
Read each field as a bounded statement
Identity. The lab applied a method it believes can tell the named substance from other things. That does not prove the method was right, or that a close cousin would have been caught.
Purity or assay. This is usually a chromatography-based estimate of how much of the signal belongs to the main peak. It is not a safety score.
Impurities. This is the “what else” question. “Not detected” is only as meaningful as the method’s ability to see what you care about.
Methods, lot, date, sample, and lab name. These tell you how and when the measurement was made. Missing or mismatched fields are already a finding.
If fields are missing, blurry, or copied from a template that does not match the label in front of you, the document has told you it is incomplete.
What a familiar test can and cannot see
Laboratories commonly use liquid chromatography to separate the parts of a dissolved sample. Software then compares the main peak with the others, and that figure is often reported as purity.
The technique is powerful when the method is sound. It can also miss what it was not set up to see, treat two similar compounds as one, or hide unknown peaks behind a tidy percentage.
Mass spectrometry is often paired with it, because a mass-related signal can support identity. Neither result is a study of what the material does in a person.
Peptides are hard to characterize
FDA has discussed, in the separate setting of drug compounding, that some peptide bulk substances raise concerns about aggregation, peptide-related impurities, and how complex the active ingredient is to characterize.
“It looks pure on this printout” and “we understand this material well” are different sentences.
On November 3, 2026, the U.S. Pharmacopeia made a related point.
Accreditation is about the lab, not about your outcome
Some COAs mention ISO/IEC 17025. That is the international reference standard for testing and calibration laboratories.
Accreditation bodies use the same standard when they assess laboratories.
None of that turns a test report into a medical green light. A competent lab can measure the wrong thing carefully, measure the right thing on an unrepresentative sample, or issue an accurate report of an irrelevant test.
A second laboratory’s COA can improve scrutiny, or repeat the same blind spots. Either way, the reports describe measured samples.
This article is general explanation. It is not a review of any seller’s files.
“Research use only” is a limitation, not a safety system
People sometimes treat “research use only” or “not for human consumption” as if it were a parallel safety system. It is not.
In FDA’s world of in vitro diagnostic products, the research-use label marks a product still in a laboratory research phase.
That context is about diagnostic products. The lesson that carries over is that the label is a limitation, not proof a product suits people.
In a February 26, 2025 warning letter about products sold online, FDA said such phrases do not, by themselves, override other evidence of intended use as a drug.
A disclaimer does not create human suitability. A COA does not finish the job the disclaimer left undone.
What a purity number can honestly say
Here is the most a purity figure can support, written as a bounded statement:
“Under the stated method, the main peak made up this share of the detector signal in the tested sample.”
That is the whole claim.
It does not, by itself, establish identity, how much of the named substance is in the vial, what the other peaks are, or whether the vial in front of you came from the tested lot.
If a seller’s page says more than that sentence, ask what additional evidence backs the extra words.
How to keep your footing
Ask what was measured, on what sample, by what method, and against what specification. Ask what was not tested, and whether the document matches the lot in front of you.
Then stop before the sentence outruns the evidence. A COA has not shown that a substance will help you, that it will not harm you, or that an unapproved product stands on the same footing as a medicine reviewed for people.
Some peptide medicines are FDA-approved drugs, available lawfully as labeled and prescribed products. They did not become medicines because a PDF looked official.
Keep that category separate from materials sold with research-use language and a certificate attached. You can respect careful laboratory work and still refuse to let a purity percentage answer a question it was never designed to answer.
This article does not recommend buying, preparing, or using research chemicals or unapproved drugs, and it gives no dosing or procurement guidance. It does not claim that anyone involved has independently verified a particular company’s certificates.
