Research
How this site grades the evidence
Every compound here carries one of three words describing the published human evidence behind the way it is sold. The word belongs to the compound and never to the seller, and the table below is generated from the same map the boards read.
Three words, and what each one means
The evidence grade is one of exactly three phrases, and the phrase carries the whole meaning.
Strong evidence means an approved indication with randomized controlled trial data behind it. Limited evidence means a real, mechanistically plausible effect with thin long-term outcome data. No controlled human evidence means little to no human trial data for the way the compound is sold.
A meter of three small squares sits beside the word, filled left to right. It is a second reading of the same fact, in ink, for anyone scanning rather than reading.
Why no color is used
A red, amber and green meter would be easier to scan and it would say something untrue.
A traffic light reads as a safety signal. Red means dangerous. This site has no basis for calling any of these compounds dangerous, and the lowest tier does not mean that. It means under-studied.
Those two claims are far apart and a color scale cannot hold them apart. So the meter is monochrome, and the word does the work.
The grade belongs to the compound, not the seller
Sermorelin's evidence grade does not change depending on who sells it. Neither does anything else's.
That is why the grade renders once near the top of a board and never beside each provider row. A grade printed in a row would read as a property of the vendor, which is the opposite of true.
A separate instrument grades the vendor. It is a numeric score built from what a provider publishes about itself, such as whether it names its compounding pharmacy and whether it publishes a price with the qualifier that makes it true. A company can score well while selling a compound with thin evidence, and both statements can appear on the same page without contradicting each other.
The grade follows the marketed use, not the molecule
This is the rule that surprises people, and it is deliberate.
PT-141 is graded no controlled human evidence here, even though the same molecule stands behind an approved product under a different name for one narrow indication in one population. The trials behind that approval were run for a use peptide storefronts do not advertise.
Grading the molecule would let an approval earned in one narrow setting vouch for claims nobody has tested. Grading the marketed use is what keeps the word honest.
The same rule applies in reverse. A compound whose parent protein has been through human trials does not inherit that record. The fragment sold under a shortened name is a different substance and is graded on its own evidence.
How a grade is actually arrived at
A grade is set by reading the published record for the compound as it is sold, and three habits do most of the work.
The first is running a positive control through the identical search filter in the same session. A search filter that returns zero for a compound is worthless unless a compound known to have hundreds of trials returns hundreds through the same filter. One widely copied version of a common trial filter silently returns zero for every term, aspirin included, and a published zero-trial count once rested on it.
The second is reading the records rather than counting them. A count answers how many papers mention a compound. It never answers whether the compound was given to anyone.
The third is splitting search terms that name different substances. Two names sitting close together in a search box merge two evidence bases into one, and the smaller of the two inherits credit it has not earned.
Three cases from this site where those habits changed the answer
Epitalon and Epithalamin are different substances: one a synthetic four-amino-acid peptide, the other a complex extracted from the pineal gland. Searched together they return five controlled trials. Split apart, all five belong to the extract and the synthetic has none.
MOTS-c returns four controlled trials. Reading all four shows every one measures how much MOTS-c a person's own body is releasing during exercise, heat stress or drug treatment. None gives MOTS-c to anyone. A count alone would have supported the opposite conclusion.
TB-500 is a seven-amino-acid fragment of thymosin beta-4, which is a protein of 43. A combined search returns seven controlled trials and every one studied the parent protein. The fragment is what is sold.
In all three cases the number was right and the thing it described was wrong. That is the failure mode the habits exist to catch.
What does not move a grade
An FDA warning letter does not move it. Most warning letters concern how a product was marketed or labeled rather than what is in a vial, and a letter discloses something worth knowing without being evidence about the medication. It never changes a grade, a score, a verdict or a ranking position.
Popularity does not move it. The most searched compound on this site sits in the lowest tier, and its search volume is not an argument about anything.
A large preclinical literature does not move it. Rodent and cell studies are where research starts, and thousands of compounds work in those models and then do not work in people.
User reports do not move it. Neither does a registered trial that has not reported. Registered means a question is being asked.
A citation count appears under a grade only where the sources support one. Where they do not, the line is left off entirely rather than printed as zero trials, because a zero would read as its own citation.
What would move a grade
Controlled human trials, in the population the compound is sold to, by the route it is sold in, with results published rather than only registered.
That is a higher bar than "some human research exists" and a lower one than "proven." It is set where it is because a buyer choosing between sellers needs to know whether anyone has tested the thing being sold, in the form being sold.
When a grade changes, the compound's board and its research page both change, and the date on each moves with it.
Every compound on this site, and its grade
16 compounds carry a grade: 2 at strong evidence, 7 at limited evidence, 7 at no controlled human evidence. This table is generated from the same map the boards and reviews read, so it cannot fall out of step with what the rest of the site says.
| Compound | Evidence grade | Board |
|---|---|---|
| Semaglutide | Strong evidence | No board |
| Tirzepatide | Strong evidence | No board |
| CJC-1295 | Limited evidence | CJC-1295 board |
| Glutathione | Limited evidence | Glutathione board |
| Ipamorelin | Limited evidence | Ipamorelin board |
| MOTS-C | Limited evidence | MOTS-C board |
| NAD+ | Limited evidence | NAD+ board |
| Sermorelin | Limited evidence | Sermorelin board |
| Tesamorelin | Limited evidence | Tesamorelin board |
| BPC-157 | No controlled human evidence | BPC-157 board |
| Epitalon | No controlled human evidence | Epitalon board |
| GHK-Cu | No controlled human evidence | GHK-Cu board |
| KPV | No controlled human evidence | KPV board |
| PT-141 | No controlled human evidence | PT-141 board |
| Semax | No controlled human evidence | Semax board |
| TB-500 | No controlled human evidence | TB-500 board |
Key takeaways
- Three phrases carry the grade: strong evidence, limited evidence, and no controlled human evidence.
- Color is never used, because a red-to-green meter would read as a safety warning rather than a research gap.
- The grade belongs to the compound and renders once per board, never beside a provider row.
- It follows the marketed use rather than the molecule, so an approval for a different indication does not transfer.
- A positive control runs through the identical search filter in the same session, or a zero count means nothing.
- A warning letter, search volume, preclinical work and user reports move no grade.
Frequently asked questions
Does a grade of no controlled human evidence mean a compound is dangerous?
No, and the wording was chosen to avoid implying it. The phrase describes what has been published, not what has been found. A compound nobody has tested in a controlled human trial is under-studied, which is a different statement from unsafe. The site uses a monochrome meter rather than a red-to-green scale for exactly this reason: a traffic light would say dangerous when the evidence only says unknown.
Why does an approved drug sometimes carry the lowest grade?
Because the grade follows the use a compound is marketed for rather than the molecule itself. An approval earned for one narrow indication in one population does not vouch for the claims made on a storefront selling the same molecule for something else. Grading the molecule would let a real trial base underwrite an untested claim.
Is the evidence grade the same as the provider score?
No. They measure different things. The evidence grade describes the published human research behind a compound and is identical no matter who sells it. The provider score describes what a company publishes about itself, such as whether it names its compounding pharmacy and whether it publishes a price with a qualifier. A well-scoring provider can sell a compound in the lowest evidence tier, and the page has to be able to say both.
Does an FDA warning letter lower a grade or a score?
No. A letter is disclosed plainly, along with what it was actually about, and the reader weighs it. The large majority of warning letters in this market concern marketing and labeling claims rather than the contents of a vial, and a letter is not a safety finding. It never moves a grade, a score, a verdict or a ranking position.
How often are grades reviewed?
They are reviewed when something changes in the published record: a trial reporting results, a new controlled study, or a regulatory decision that changes what is being sold. Each compound page and each board carries the date of its last update, so a grade that has not been re-read is visible as an old date rather than presented as current.
Sources
Each document below is named as it names itself, with the date printed on that document rather than the day it was read.
- FDA Briefing Document for Epitalon-Related Bulk Drug Substances, Pharmacy Compounding Advisory Committee — U.S. Food and Drug Administration, May 2026
- FDA Briefing Document for MOTS-c-Related Bulk Drug Substances, Pharmacy Compounding Advisory Committee — U.S. Food and Drug Administration, May 2026
- Certain Bulk Drug Substances for Use in Compounding that May Present Significant Safety Risks — U.S. Food and Drug Administration, April 2026
- PubMed literature censuses for this batch, each run with a positive control through the same filter — National Library of Medicine, September 2026
- Registered study censuses for this batch, each run with a positive control through the same query — ClinicalTrials.gov, National Library of Medicine, September 2026