Every Kovera Labs COA, searchable.

A match here shows what the lab's record said on 2 August 2026. Kovera has revised 57certificates under the same number since first publication, so open the lab's verify page or run the code through our COA verifier before you rely on a figure.

1,828 certificates analysed between 11 Jan 2026 and 27 Jul 2026. The lab graded 1,742 as a pass and 17 as a fail; 69 carry no verdict. 121 client names appear after merging spellings that differ only in capitals or spacing. A further 117 certificates name the client only by an email address, which we do not republish, and 91 print no client.

24 certificates meet at least one of our failure rules: a lab fail, an identity the lab did not confirm, purity under 95%, or measured content under 90% of the label. Another 861 measured more than 110% of the label, which the lab does not grade as a fail. The rules and the full list are on the Kovera failures page.

We publish a page for each of the 76 clients with five or more certificates. The whole index is a free CSV: kovera-certificates-2026-08-02.csv (listed on our data page).

Each row is a certificate as Kovera published it: the client name the lab printed, the product name on the certificate, the lot the client supplied, HPLC purity, and measured content against the label claim. We read each certificate into those fields; a misread digit is possible, so the lab's verify page is the reference.

Clients choose which products to send and whether a certificate goes public, so this is the record vendors chose to publish, not a random sample of what ships. A certificate covers the sample tested, not every vial sold under the same lot. Where a client name is an email address we show it as withheld. We have not ordered from any client listed and have sent no sample to Kovera.

Content over 110% of the label usually means an overfilled vial. On blends, the label often states the amount of each component while the lab reports the total, which can push the percentage near 200%; read the certificate before treating that figure as a finding.