Peptide COA: How to Read and Verify a Certificate of Analysis

Bottom line: A peptide COA (certificate of analysis) names the laboratory and the sample, often with a batch code, dates the work, lists the tests requested and gives the results, usually a purity percentage and sometimes the measured amount of peptide. To check one, open the report on the laboratory’s own site rather than the seller’s copy, match the report number, compound, results, batch and date, and read which tests were run. If all five checks pass, the certificate supports what the lab measured in one submitted sample. It does not prove that every vial in the seller’s inventory came from that sample, or that untested attributes such as sterility are acceptable.

How to read and verify a peptide COA: five checks

A QR code is a shortcut, not evidence. Start the lookup yourself whenever possible: type the report number into the lab’s site or use TitrateLab’s peptide COA verification tool. A QR code that points back to the seller, a PDF uploaded only to the seller’s domain, or a screenshot of a laboratory page leaves the seller in control of the evidence.

The worked example below is a public Janoshik report on a semax sample. Markers 1 to 5 point to the field each of the five checks reads; marker 6 is the results table, discussed under purity.

1. Verify the report on the laboratory’s server

The strongest basic authenticity check is a record controlled by the organization that issued the result. Janoshik maintains a public test database; other laboratories and testing programs use their own report pages, accessions or verification keys. On the example, marker 1 points to the task number, #171494, and to the unique key at the foot, HZ7NP461XI5W, printed under a line telling readers to verify the test at janoshik.com/verify. Janoshik’s own page for that key should show the same task number and results as the copy you were sent.

This step can catch a fabricated document, an altered purity number or a report number that does not exist. It cannot establish that the tested sample came from the seller presenting it.

2. Match every material field

Do not stop when the report number resolves. Compare:

• the compound or sample name;
• the stated quantity or vial size;
• the purity and measured amount;
• the test date;
• the named client or submitter, when disclosed; and
• the laboratory’s report number, key or accession.

A genuine report can still be mis-linked. If the laboratory record names a different compound, quantity or client than the seller’s page, the certificate may be authentic while the seller’s claim is unsupported.

Marker 2 on the example covers the client, the sample name (“semax 10mg”) and the manufacturer. The client and the manufacturer are the same business: utherpeptide.com, the storefront of the supplier UTHER. As printed, the supplier ordered a test of its own product, so it also chose which vial to send: the selected-sample case described further down.

3. Match the batch to the vial

The batch or lot identifier is the bridge between a laboratory result and a physical product. It should appear on the report and on the vial, packaging or batch-specific documentation supplied with it. A report without a batch identifier may still be a real assay, but it cannot perform this traceability job by itself.

The example’s batch (marker 3) is SX10-0526. Our comparison of domestic and overseas sellers found a domestic seller, Porter Peptides, printing that same code on its own semax label. A second laboratory, Freedom Diagnostics, reported on that lot for the client Porter Peptides; its report, below, photographs the labelled vials and prints the lot as SX10-0526. One code then appears on Porter’s vial, on Porter’s report and on UTHER’s own report. That points to a supply connection. It does not prove the vials came from one production run, and it depends on the code being printed honestly. To look for public reports that print your code, search Find Your Batch by compound and code.

In TitrateLab’s August 10 snapshot, 52,163 of 73,865 published records (70.6%) carried a batch identifier. The remaining 21,702 did not. Absence does not prove anything was falsified; it means the record cannot be joined to a labelled batch using that field.

Some missing values are extraction gaps on our side. The Janoshik report prints its batch, but on October 11, 2026 our record of it had no batch value. A Find Your Batch search for semax and SX10-0526 that day checked 1,702 semax reports and returned one exact match: Porter’s Freedom Diagnostics report above. A batch search can only match codes that were captured, so one result, or none, does not show that no other report exists. Search the laboratory’s own database as well.

4. Check the date and inventory continuity

A certificate is dated evidence. A passing test from 2024 does not establish that a differently numbered batch sold in 2026 matches it. Even a current report can be too early if it tested a pre-production sample rather than stock selected from the finished batch.

Ask a simple question: what connects the tested sample, the labelled batch and the vial being represented today? A matching lot number is useful. Independent sampling from finished inventory is stronger. A generic “tested by” badge is not a chain of custody.

Marker 4 on the example shows three dates: testing ordered June 2, sample received June 5 and analysis conducted June 15, 2026. The report covers that sample on those dates, not vials sold later under the same code.

5. Read the test scope

“Lab tested” is incomplete without the next words: tested for what? Identity, purity, quantity, endotoxin, sterility, residual solvents and heavy metals are different measurements. One result cannot silently stand in for the others.

In the Janoshik example (marker 5), the only test requested was “Semax analysis”, and it returned an amount and a purity without naming a method. Porter’s Freedom Diagnostics report labels its result lines by method: identity by LC-MS, purity by HPLC-UV, and a separate net-content line. Neither report addresses endotoxin, sterility, residual solvents or heavy metals, so both leave them unknown.

Janoshik lists peptide composition testing and endotoxin analysis as separate services. Finnrick’s published testing methodology separately evaluates identity, purity, fill accuracy and batch labeling. That separation is the correct way to read a certificate: line by line, not as a universal pass badge.

If the report you hold does not test what you need, choose a lab to test your own vial. Our guide to getting a peptide tested explains which test answers which question and how many vials to send.

What a valid COA actually proves

A verified certificate supports a narrow statement:

The issuing laboratory reports that it received a sample and obtained the listed results using the stated analyses.

That is useful evidence. It is materially stronger than a seller’s promise or an in-house spreadsheet. But the word valid is often stretched far beyond that statement.

A lab-hosted record does not, by itself, establish:

• who selected the submitted sample;
• whether it was randomly drawn from finished inventory;
• whether the seller presenting the report commissioned or owns it;
• whether the vial in front of you carries the same batch;
• whether later batches perform the same way;
• whether another laboratory would report the same result; or
• whether attributes absent from the report were tested and passed.

Two reports on one printed code can disagree. For semax lot SX10-0526, Janoshik measured 11.92 mg for the client utherpeptide.com and Freedom Diagnostics measured 13.64 mg for the client Porter Peptides, against 10 mg stated on each report. The 1.72 mg gap may come from the labs, from the vials or from both; two reports cannot separate them. Across the market, our August 27, 2026 review of 179 carefully matched batch-code groups tested by more than one lab found a major conflict in 123: in 118 the measured amount differed by at least 10% of the label, and in five the labs disagreed about which peptide was present. Each of those groups held one vendor’s reports, so this two-seller pair is not among them, although its gap of 17.2% of the label is above the review’s 10% line. Those records show disagreement, not which lab was right. The full review explains why peptide labs report different results for the same batch code.

Regulated manufacturing guidance illustrates how much infrastructure normally sits around a COA. FDA’s Q7 guidance for active pharmaceutical ingredients calls for batch identification, the tests performed, acceptance limits and numerical results, while also requiring manufacturers to evaluate suppliers and periodically check the reliability of their certificates. FDA’s component-testing guidance likewise emphasizes representative lot sampling and independent identity testing when a manufacturer relies on supplier analysis.

Grey-market research-peptide certificates are not evidence that those pharmaceutical controls exist. The comparison matters because it shows what a PDF cannot replace: qualified suppliers, representative sampling, batch records, specifications, review and continuing verification.

Why “99% purity” may leave the main question unanswered

The most visible number on a peptide COA is usually purity. It is also one of the easiest numbers to overread.

HPLC purity describes the relative composition detected in an assay. It does not answer how many milligrams of active compound are in the vial. A small amount of highly pure material can produce an excellent purity percentage while the container remains substantially short of its labelled amount. Quantity testing is the measurement that compares the active amount found with the label claim.

The Janoshik example shows both lines (marker 6): 11.92 mg of semax in a sample submitted as “semax 10mg”, 19.2% above the stated amount, and a purity of 99.020%. The purity line could not have shown that difference; only the amount line did. A gap between measured and labelled amount carries into every concentration worked out from the label: mixed with the same volume of water, 11.92 mg makes a solution 19.2% stronger than 10 mg would. The peptide reconstitution calculator shows that arithmetic for any vial amount and water volume.

The coverage gap is visible in TitrateLab’s current corpus:

The correct reading is not that the other certificates failed quantity or endotoxin testing. It is that the relevant field was absent or not usable in this dataset. Not tested, not recorded and failed are three different states. Collapsing them into one green “verified” badge destroys the information a certificate is supposed to preserve.

The same distinction applies to sterility. An HPLC purity result is not a sterility result, and an endotoxin result is not a sterility result. Each test has its own method, sample requirements and evidence boundary.

Our retatrutide bait-and-switch advisory shows why identity comes first: one sample labelled retatrutide reported tirzepatide at 99.269% purity. It includes the source report and practical checks.

Four ways a genuine certificate can still mislead

The borrowed report

A seller links to a genuine laboratory result commissioned by another company or for another supply chain. The document is authentic; the relationship to the seller’s vial is not established. Compare the disclosed client, product details and batch rather than treating the lab’s logo as transferable trust.

The recycled report

One successful batch is used to represent later inventory. Check the lot and date. A report can remain genuine for years while becoming irrelevant to the stock being sold today.

The selected sample

The vendor or manufacturer chooses which vial goes to the laboratory. That result describes the chosen sample. It cannot estimate the failure rate of inventory that was never sampled, especially when unsuccessful private tests may never be published. Our guide to checking whether a peptide vendor is legit shows how to find results for vials the seller did not choose.

The incomplete panel

The certificate reports a strong purity result and omits quantity, endotoxin or sterility. Nothing on the document is false; the mistake happens when “passed purity” is translated into “passed everything”. Read absent tests as unknown.

A practical evidence ladder

Not all certificates deserve equal weight. This is the order TitrateLab uses when thinking about decision value:

1. Seller-hosted image or PDF: a claim that may be genuine, but the issuer has not yet been checked.
2. Laboratory-hosted matching record: stronger authenticity; the reported values can be compared with the seller’s copy.
3. Recent, batch-matched laboratory record: adds traceability from the report to a labelled batch.
4. Identity, purity and quantity on the same relevant sample: answers three separate product questions instead of displaying purity alone.
5. Independently selected finished-product sample with appropriate additional testing: reduces seller-selection bias and broadens the attributes measured.

Moving up the ladder does not convert an unregulated product into an approved medicine or guarantee every vial. It makes the evidence narrower, clearer and harder to misrepresent.

Verify a report in TitrateLab

Use the COA verification lookup when you have a Janoshik test number or report identifier; it does not search batch codes. Without Pro, it tells you whether the number is in TitrateLab’s certificate archive; the archived peptide, purity, amount, lab, batch, date, source document and certificate image are part of Pro. On October 11, 2026, a lookup of 171494, the Janoshik example above, returned a match. When you have a printed lot code but no report number, Find Your Batch searches public reports by compound and code. You can also search the broader COA database by compound and inspect TitrateLab’s methodology for how records are attributed and normalized.

When reviewing a result, write down five answers:

1. Where is the laboratory-hosted record?
2. Do its material fields match the seller’s claim?
3. Does its batch match the vial or current listing?
4. Is the test recent enough to be relevant?
5. Which important attributes remain untested or unrecorded?

If the answer to the last question is “none”, check again. No single peptide certificate ordinarily answers every identity, composition, dose, microbiological and contamination question.

Method, evidence boundary and corrections

TitrateLab queried the current certificate corpus read-only at 2026-08-10 21:31 UTC. The snapshot contained 73,865 published records from 15 ingest sources, with test dates through August 9, 2026. The accompanying how-to-verify-a-peptide-coa.queries.sql file contains the exact coverage queries used for the figures above and the record query for the worked example.

Counts are certificate records, not a random sample of the market and not a count of unique vials. Public evidence is selected: vendors, manufacturers, testing groups and buyers choose what to test and what to publish. A populated field shows that TitrateLab recorded a value; it does not independently validate the laboratory method or prove that a submitted sample represents unsampled inventory. An empty field can also mean our extraction missed a value the document prints, as the batch example above shows. Lab names can also be anonymized or normalized from source documents.

The annotated certificate is Janoshik report #171494, rendered from TitrateLab’s archived copy of the public report, saved July 11, 2026. A direct request to the report’s page on October 11, 2026 was stopped by the site’s bot check, so the figure uses the archived copy. The Freedom Diagnostics report was downloaded from the laboratory’s report server on October 11, 2026. The verification lookup, the Find Your Batch result and our record’s missing batch value were observed the same day. The cross-lab figures come from our August 27, 2026 review.

The interpretive standard comes from primary laboratory documentation and FDA quality guidance. FDA requirements are included as a regulated benchmark, not as a claim that grey-market sellers or community testing programs operate under pharmaceutical CGMP controls.

This article is research and education, not medical advice, a product endorsement or a guarantee of identity, quality, legality or safety. If a source record or calculation is wrong, use the correction link in the footer; TitrateLab updates the evidence trail rather than silently changing the conclusion.

Open original ×