Data for journalists and researchers.
The tables behind our reporting on FDA warning letters and peptide lab results, as dated CSV files you can download, check and reuse. Each file links to the article it comes from, and every row links to the public record it was built from. Credit vialaudit and link to the source page.
The datasets
DATED SNAPSHOTS · UTF-8 CSV| Dataset | Rows | Snapshot | File |
|---|---|---|---|
| Retatrutide sellers after an FDA letter or Lilly lawsuit | 33 | 10 October 2026 | still-selling-2026-10-10.csv · 10 KB |
| FDA warning letters that cite retatrutide | 22 | 10 October 2026 | fda-retatrutide-letters.csv · 4 KB |
| Janoshik public retatrutide and tirzepatide reports | 116 | 30 June 2026 | janoshik-retatrutide-tirzepatide-2026-06-30.csv · 21 KB |
| Kovera Labs certificates per vendor | 124 | 2 August 2026 | kovera-vendor-summary-2026-08.csv · 7 KB |
| BPC-157 + TB-500 ("Wolverine") blend lab results | 81 | 2 August 2026 (Kovera) and 30 June 2026 (Janoshik) | wolverine-blends-2026.csv · 13 KB |
Files use a header row, commas, double-quoted text where needed, ISO 8601 dates and no comment lines. A blank cell means the source did not give that value. Filenames carry the snapshot date, so a link to a file keeps pointing at the same snapshot.
Retatrutide sellers after an FDA letter or Lilly lawsuit
33 ROWS · 2026-10-10One row for each of the 22 FDA warning letters that cite retatrutide and each of Eli Lilly's 11 retatrutide lawsuits, with what the named website showed when we loaded it on 10 October 2026.
Source page: FDA retatrutide warning letters: who still sells it
How we built it
We loaded every domain named in a letter or complaint between 05:22 and 05:36 UTC on 10 October 2026, then loaded the listed and login-walled sites again between 05:45 and 05:50 UTC. Where a complaint was not public we used the domain matching the business name and marked it "inferred".
"Listed" means a product page naming retatrutide or a code name loaded with a price and a cart button. Legendary Peptides is the one exception: we saw the product and price once but could not confirm a cart button.
The status_group column reproduces the summary table in the source article: 6 FDA-letter sites and 2 Lilly defendants listed on the named domain, 1 more on a redirect target whose operator we have not confirmed.
Limits. A listing shows what a page offered that morning. It is not proof of a sale, and none of it is a finding that any seller broke the law. We bought nothing and created no accounts.
Columns
| Column | Meaning |
|---|---|
| recipient | Company as named in the FDA letter or Lilly complaint |
| action_type | FDA warning letter or Eli Lilly lawsuit |
| letter_no_or_case_no | FDA letter number, or court and case number |
| action_date | Letter date or filing date, ISO 8601 |
| domain_checked | Website we loaded; "inferred" where the complaint named none |
| site_status_2026_10_10 | HTTP result or what the site showed |
| listed | yes, no, could not check, or not checked |
| listing_detail | Product name, stock flag and cart button as we saw them |
| status_group | The summary-table bucket the row counts toward |
| notes | Redirects, renames, registry dates |
| source_url | FDA letter or court filing, where public |
Suggested citation
vialaudit, "Retatrutide sellers after an FDA letter or Lilly lawsuit", dataset, snapshot 10 October 2026. https://vialaudit.com/data/still-selling-2026-10-10.csv. CC BY 4.0.FDA warning letters that cite retatrutide
22 ROWS · 2026-10-10The 22 letters from FDA's drug center that cite retatrutide, issued 10 December 2024 to 24 August 2026. Twenty-one went to sellers; one went to an ingredient distributor.
Source page: FDA retatrutide warning letters: who still sells it
How we built it
We pulled FDA's warning-letter index on 10 October 2026, kept every drug-center (CDER) letter issued since 1 November 2024 (583 letters), read the full text and searched for retatrutide, LY3437943 and the code names FDA quotes ("GLP-3 RT", "GLP-3R", "GLP-1-R", "GLP1-R", "RETA").
domain_named is the website the letter says FDA reviewed. It is blank for GenoGenix, whose letter followed an inspection and names no site.
Limits. A letter that names the drug only inside an image or attachment would not show up in a text search. FDA's letters state FDA's position; recipients may dispute it.
Columns
| Column | Meaning |
|---|---|
| letter_no | FDA's reference number for the letter |
| letter_date | Issue date, ISO 8601 |
| recipient_company | Company as named by FDA |
| domain_named | Website FDA reviewed, if any |
| letter_url | The letter on fda.gov |
Suggested citation
vialaudit, "FDA warning letters that cite retatrutide", dataset, snapshot 10 October 2026. https://vialaudit.com/data/fda-retatrutide-letters.csv. CC BY 4.0.Janoshik public retatrutide and tirzepatide reports
116 ROWS · 2026-06-30All 116 retatrutide (68) and tirzepatide (48) reports in our 30 June 2026 mirror of Janoshik Analytical's public test listing, with label size, measured content and purity.
Source page: Janoshik tirzepatide and retatrutide results: 116 reports
How we built it
We read each public report image into structured fields. pct_of_label is measured_mg divided by the mg printed in the sample name. Thirteen reports have no pct_of_label, and pct_of_label_missing_reason says why: no label size, a pen reported in mg/ml, a qualitative LC-MS screen, or no mg figure.
vendor_as_printed is the client name on the report. Where the client field was an email address we show the company website printed on the report instead, never the address.
The medians in the source article (110.4% and 112.7% of label; 99.78% and 99.81% purity) recompute from this file.
Limits. Vendors choose which reports go public, so this is the best case the market shows, not a random sample. Twenty reports share identical figures with another report, which points to one tested sample per group. A misread digit is possible outside the outliers we re-checked.
Columns
| Column | Meaning |
|---|---|
| report_id | Janoshik task number |
| compound | Retatrutide or Tirzepatide |
| sample_as_printed | Sample name on the report |
| label_mg | mg in the sample name |
| measured_mg | mg Janoshik measured |
| pct_of_label | measured_mg / label_mg x 100, one decimal |
| purity_pct | HPLC purity, % |
| identity_confirmed | yes if Janoshik confirmed the compound |
| test_date | Analysis date, ISO 8601 |
| vendor_as_printed | Client name or company website on the report |
| pct_of_label_missing_reason | Why pct_of_label is blank, if it is |
| report_url | The report on verify.janoshik.com |
Suggested citation
vialaudit, "Janoshik public retatrutide and tirzepatide reports", dataset, snapshot 30 June 2026. https://vialaudit.com/data/janoshik-retatrutide-tirzepatide-2026-06-30.csv. CC BY 4.0.Kovera Labs certificates per vendor
124 ROWS · 2026-08-02One row per named client in our mirror of 1,828 Kovera Labs certificates retrieved 2 August 2026: 124 names covering 1,620 certificates.
Source page: Kovera Labs: what 1,828 published COAs show
How we built it
We left out 208 certificates: 91 print no client name and 117 name the client only by an email address. We do not republish email addresses.
Names are as printed. We did not merge spellings, so "RemedyX Labs" and "REMEDYX LABS" are separate rows.
explicit_fails counts certificates Kovera marked as failed; identity_mismatches counts those where the identity check failed; below_95_purity counts purity under 95%. The median is over certificates that report a purity.
Limits. This file uses the full August mirror, so a few counts differ from our May leaderboard, which used the 26 May snapshot of 1,816 certificates (Ion Peptide has 106 here, 105 there). Kovera revises some certificates in place; values are as retrieved on 2 August 2026.
Columns
| Column | Meaning |
|---|---|
| vendor_as_printed | Client name on the certificates |
| certificates | Certificates in the mirror |
| certificates_with_purity | Certificates that report an HPLC purity |
| median_purity_pct | Median purity, % |
| lowest_purity_pct | Lowest purity, % |
| explicit_fails | Certificates marked failed |
| identity_mismatches | Certificates with a failed identity check |
| below_95_purity | Certificates under 95% purity |
| first_test_date | Earliest analysis date |
| last_test_date | Latest analysis date |
Suggested citation
vialaudit, "Kovera Labs certificates per vendor", dataset, snapshot 2 August 2026. https://vialaudit.com/data/kovera-vendor-summary-2026-08.csv. CC BY 4.0.BPC-157 + TB-500 ("Wolverine") blend lab results
81 ROWS · 2026-08-0281 public lab results for BPC-157 plus TB-500 blend vials: 69 Kovera Labs certificates (February to May 2026) and 12 Janoshik reports from 10 sellers (April 2025 to June 2026).
Source page: Wolverine stack: BPC-157 + TB-500 evidence and lab tests
How we built it
Kovera rows are every certificate for a BPC-157/TB-500 blend, including six labelled "Wolverine". Kovera reports one purity and one net content per vial. Four rows read about 200% of label, most likely because the label gave the amount per peptide.
Janoshik rows report each peptide's mg separately and no purity. Janoshik does not grade these reports, so lab_result reads "not graded". Report 83375 gives the mean of two runs; for report 172940, which lists two replicates, we show the higher value of each.
vendor_as_printed is blank where Kovera printed no client name or only an email address.
Limits. A blend certificate does not show the ratio of the two peptides or which molecule a seller calls TB-500. The one Kovera failure (KVR-2026-550573) was on fill, 11.2% under label.
Columns
| Column | Meaning |
|---|---|
| source | Kovera Labs or Janoshik |
| certificate_or_report_id | Kovera certificate number or Janoshik task number |
| product_as_labelled | Product name on the certificate |
| vendor_as_printed | Client name on the certificate |
| test_date | Analysis date, ISO 8601 |
| label_mg | Label mg (Kovera); per-peptide label for Janoshik |
| measured_mg | Net content measured by Kovera |
| pct_of_label | Kovera measured / label x 100 |
| purity_pct | Kovera HPLC purity, % |
| bpc157_mg, tb500_mg | Janoshik per-peptide mg |
| bpc157_pct_of_label, tb500_pct_of_label | Janoshik per-peptide % of label |
| lab_result | pass, fail, or not graded |
| verify_url | The record on the lab's own verify page |
Suggested citation
vialaudit, "BPC-157 + TB-500 ("Wolverine") blend lab results", dataset, snapshot 2 August 2026 (Kovera) and 30 June 2026 (Janoshik). https://vialaudit.com/data/wolverine-blends-2026.csv. CC BY 4.0.Charts you can embed
PNG · 1600 × 1000Each chart is a static image built from the files above. Embed it with the snippet under it, which keeps the credit and the link back to the source article. Please do not crop out the source line.
<figure>
<a href="https://vialaudit.com/articles/still-selling-after-fda-letter">
<img src="https://vialaudit.com/images/charts/still-selling-status-2026-10-10.png" alt="Bar chart: of 22 FDA retatrutide warning-letter websites on 10 October 2026, 6 still listed retatrutide on the named domain, 1 on a redirect target, 6 did not, 3 were gone, 5 could not be checked and 1 named no website. Of 11 Lilly lawsuit defendants, 2 listed it." width="1600" height="1000" loading="lazy">
</a>
<figcaption>Chart: <a href="https://vialaudit.com/articles/still-selling-after-fda-letter">vialaudit</a>, CC BY 4.0</figcaption>
</figure>
<figure>
<a href="https://vialaudit.com/articles/janoshik-tirzepatide-retatrutide-results">
<img src="https://vialaudit.com/images/charts/janoshik-reta-tirz-pct-of-label-2026-06-30.png" alt="Two histograms of measured content as a percentage of label from Janoshik public reports to 30 June 2026. Retatrutide: 60 reports, median 110.4%. Tirzepatide: 43 reports, median 112.7%. Most vials sit above 100%." width="1600" height="1000" loading="lazy">
</a>
<figcaption>Chart: <a href="https://vialaudit.com/articles/janoshik-tirzepatide-retatrutide-results">vialaudit</a>, CC BY 4.0</figcaption>
</figure>
<figure>
<a href="https://vialaudit.com/articles/kovera-1816-coas-vendor-analysis">
<img src="https://vialaudit.com/images/charts/kovera-top-15-vendors-2026-08.png" alt="Horizontal bar chart of Kovera Labs certificates per vendor, 2 August 2026. Ion Peptide leads with 106. Glacier Aminos (78), Instant Peptides (72) and Peptide Partners (61) are highlighted for at least one failed or identity-mismatch certificate." width="1600" height="1000" loading="lazy">
</a>
<figcaption>Chart: <a href="https://vialaudit.com/articles/kovera-1816-coas-vendor-analysis">vialaudit</a>, CC BY 4.0</figcaption>
</figure>
<figure>
<a href="https://vialaudit.com/articles/bpc-157-tb-500-stack">
<img src="https://vialaudit.com/images/charts/wolverine-blend-purity-2026.png" alt="Histogram of HPLC purity for 69 Kovera Labs BPC-157 plus TB-500 blend certificates, February to May 2026. All sit between 98.7% and 100%, median 99.76%." width="1600" height="1000" loading="lazy">
</a>
<figcaption>Chart: <a href="https://vialaudit.com/articles/bpc-157-tb-500-stack">vialaudit</a>, CC BY 4.0</figcaption>
</figure>Licence and how to credit us
CC BY 4.0You can copy, chart, quote and republish these files and charts, including in commercial work, under the Creative Commons Attribution 4.0 licence. The one condition is credit: name vialaudit and link to the source page listed for the dataset, or to this page.
A short credit line works: Source: vialaudit (vialaudit.com/data), snapshot date. Please give the snapshot date. These files describe what public records and websites showed on a given day, and some of them have changed since.
The licence covers our compilation and the fields we extracted. It does not cover the underlying certificates, reports and letters, which belong to the labs and agencies that issued them. Each row links to the original so you can check it yourself.
Press contact and corrections
INFO@VIALAUDIT.COMFor background, a fact-check or a question about a row, email info@vialaudit.com. Our press kit has quotable figures and story angles.
If you find an error in a file, send us the filename and the row. We correct the file at the same address and record the change, with its date, in our corrections log. How we test and what we publish is set out in our methodology.
Independence and disclosure
0 SPONSORED ROWSWe earn commission on some vendor links elsewhere on the site. Audits, scores and rankings are independent, and vendors do not pay for placement. No file on this page contains an affiliate link. The full policy is on our affiliate disclosure page.
All vials referenced are research-grade reagents and are not intended for human consumption. Nothing in these files is dosing or medical guidance.