BRAK LABS Peptide research, plainly written

question

Do peptides work?

The evidence sits per compound and per claim, which is why every claim on this site carries a tier. Three worked examples show what that looks like in practice.

Last Reviewed Editorial policy Methodology

Peptide names a class of chemistry covering hundreds of molecules, each with its own target, its own published record and its own amount of testing behind it. So the useful question is always about one compound and one claim, and this page shows how to ask it.

This page shows what that looks like. Three compounds from this site's own reviews, each in a different position in the published record, and the reading each one supports.

Not medical advice

This page reports what has been documented about a compound. It is not medical advice, not a diagnosis, not a dosing protocol, and not a recommendation to obtain or use anything described here. Talk to a licensed clinician about anything concerning your health. Read the full disclaimer.

Reviewed against editorial standards Updated

The short answer

Evidence exists per compound, and within a compound it exists per claim. One molecule can have a large randomised trial behind one claim and nothing but forum accounts behind the next, so a single verdict across the category would be false in both directions at once.

That is the reason this site runs two devices instead of a verdict. Every claim about what a compound does carries one of five evidence tiers, from Human RCT down to Anecdote, assigned from the strongest source that actually tested that specific claim. Every compound with a published review carries a grade derived from those tiers, and never asserted on its own. The methodology page sets out both.

The two questions that decide any answer

Reading a claim about any of these compounds comes down to two questions, in this order.

First, was this compound tested, or was something related to it tested? A peptide that is a fragment of a larger protein does not inherit that protein's trials, and neither does a close analog or a different salt. This is the commonest way a compound in this category acquires a trial count it never earned, and every grade on this site is derived with that rule applied.

Second, did the testing measure the thing being claimed? A completed trial on wound healing says nothing about a cognition claim. A study that measures a hormone concentration in blood has measured a hormone concentration in blood, which is a different thing from measuring what a person notices.

The three examples below are chosen because each one answers those two questions differently.

Example one: the most tested compound on this site

Thymosin alpha-1 has more completed human trials behind it than anything else covered here, and it is registered as a medicine outside the United States. It is the compound where the answer to "has this been tested in people" is unambiguously yes.

Its most rigorously tested claim was asked twice at scale. A 2013 multicentre randomised trial in 361 patients with severe sepsis reported 28 day mortality of 26.0 per cent against 35.0 per cent [2]. A 2025 phase 3 trial in 1,106 adults, quadruple-masked against placebo across 22 centres, reported 23.4 per cent against 24.1 per cent, hazard ratio 0.99, 95 per cent confidence interval 0.77 to 1.27, P equal to 0.93 [Human RCT] [1] [2]. The larger trial is roughly three times the size, placebo-controlled, and masked from participants, carers, investigators and outcome assessors, and its authors state that no secondary or safety outcome differed statistically significantly.

So the compound with the most human data on this site has a null result from its biggest and best-designed trial. That is worth sitting with, because it is the opposite of what a category-level answer would predict: more testing produced a clearer negative rather than a clearer positive. It is also why this site's grades measure how much research exists and how good it is, and say nothing about which way it came out. Direction travels with the explanation beside the grade, never inside the letter.

The full review sets out six claim areas, each with its own tier, including the ones where the direction is favourable.

Example two: a compound with no human record at all

KPV sits at the other end. It is a three amino acid fragment, and no human study of any design has tested it on any route. FDA reached the same finding independently in its own briefing document of 12 May 2026, searching PubMed, Embase, DailyMed, Drugs@FDA and the adverse event reporting system [8].

What exists instead is rodent work. Two of those studies are cited here, both in mouse models of chemically induced inflammatory bowel disease, measuring the inflammation in the colon and the animals' recovery [Animal] [3] [4]. Work of that kind supports a measurable difference in how a chemically injured mouse colon recovers over days under laboratory conditions. It does not show an effect on inflammatory bowel disease in a person, on any route. The KPV review sets out the whole set, each study with its own citation.

The distance between those two sentences is the entire content of the question this page is answering. It is not a criticism of the studies, which are ordinary preclinical work of the kind every drug programme starts with. It is a statement about what preclinical work can carry.

What this evidence can and cannot show

These results come from mice in chemically induced models of inflammatory bowel disease, not from people. An animal model is a deliberate simplification: the injury is created on purpose, the animal is young and healthy, the dose is scaled to body weight in a way that does not translate directly, and the outcome is measured at a fixed point rather than lived with. Cell and tissue work is a further step removed, because the concentrations that produce an effect in a dish are often far above anything a body reaches. Mechanism is a reason to run a trial. It is not a substitute for one, and compounds that looked mechanistically convincing have failed in people many times.

Example three: one name, two products, two different answers

The sharpest illustration on this site is a pair sold under the same name. CJC-1295 is a growth hormone releasing hormone analogue, and it is sold in two forms: one carrying an albumin-binding linker called a Drug Affinity Complex, and one without it. The base peptide is identical in both.

The version with the linker has a published human record. Two randomised, placebo-controlled, double-blind ascending studies in healthy adults aged 21 to 61 reported mean plasma growth hormone rising 2 to 10 fold for six days or more after a single administration, mean IGF-1 rising 1.5 to 3 fold for nine to eleven days, and an estimated half-life of 5.8 to 8.1 days [5]. A second study sampled blood every 20 minutes across a 12 hour overnight window in healthy men and found growth hormone pulse frequency and magnitude unaltered, with the trough between pulses rising 7.5 fold [Human RCT] [5] [6]. Every one of those measurements is a concentration in blood. No study of that compound has measured body composition, strength, recovery or sleep.

The version sold without the linker has no published study under its own name. A PubMed search on 23 August 2026 returned three published human studies carrying the CJC-1295 name, and all three used the version with the linker, so the half-life and duration figures quoted for the version without it trace to community and vendor writing rather than to a study of that product [Anecdote] [7].

Two products, one chemical linker apart, at opposite ends of the evidence scale. Nothing measured on one is stated of the other anywhere on this site, and the page that separates them exists because the market routinely states it.

What would move any of these answers

A tier moves up when a qualifying study is published, cited by number, and added to the trials table on that compound's page in the same edit. It never moves because a claim became more popular.

That means the answer for a given compound is dated, and every page here carries a visible Last Reviewed stamp saying when a person went through it. For most of the compounds covered, the thing that would change the answer is the first completed human trial, and for several of them nothing is currently registered that would produce one.

Where a tier changes because the previous one was wrong rather than because new evidence landed, the change is written up in the corrections log with both readings stated. That distinction is recorded in the correction itself.

How to check any claim on this site in about a minute

Open the compound's page and go to its Claim Strength Matrix, which sits near the top. Every claim area the compound is discussed for has a row, including the weak ones, and each row carries the tier of the work behind it and a column stating what that work can and cannot show.

The bracketed chip beside a claim links to the tier's own definition on the methodology page, so the vocabulary is checkable from anywhere. The bracketed numbers beside it point into that page's numbered reference list, where every entry carries a PubMed identifier, a digital object identifier, or the agency document it came from.

The compound index lists everything covered, grouped by what each one is studied for.

Our takeThe compound on this site with the most human data behind it has a null result from its largest trial, and the compound with no human data at all has a shelf of rodent work. Any single answer for the category would misdescribe both.

We report what has been documented. We do not prescribe what should be done.

Frequently asked questions

Do peptides work?

There is no answer at the level of the category, because peptide names a class of chemistry rather than a treatment. Evidence exists per compound and per claim: one molecule can have a large randomised trial behind one claim and forum accounts behind the next. Every claim on this site carries a tier recording what it rests on.

Which peptide has the strongest evidence on this site?

Thymosin alpha-1 has the most completed human trials behind it, including a 2025 phase 3 trial in 1,106 adults. That trial reported no statistically significant difference from placebo on its primary outcome, which is a good illustration of why the amount of evidence and the direction of the result are two separate things.

If a peptide has animal studies behind it, does that mean it will work in people?

No. An animal model is a deliberate simplification: the injury is created on purpose, the animal is young and healthy, the amount given is scaled to body weight in a way that does not translate directly, and the outcome is measured at a fixed point. Mechanism is a reason to run a trial rather than a substitute for one, and compounds that looked mechanistically convincing have failed in people many times.

Why does this site not just say whether a compound is worth trying?

Because that is a recommendation, and this site reports what has been documented rather than prescribing what should be done. The people who publish it also have a commercial interest in the sale of peptide products, which the editorial policy states in advance. This site links to no seller today, and when it does, every such link will carry a visible disclosure of that relationship beside it. Either way a recommendation here would not be independent, so none is made.

Does a higher evidence grade mean better results?

No. A grade measures how much research exists about a compound and how good that research is. It records nothing about which way the results went. A compound tested twice against a control and beaten twice still reaches a higher letter than one nobody has ever tested, and the explanation beside the grade is where that direction is stated.

Two vendors sell the same peptide under slightly different names. Is the evidence the same?

Not necessarily, and CJC-1295 is the worked example on this page. The version carrying an albumin-binding linker has three published human studies; the version sold without that linker has none under its own name. One chemical difference separates them, and the published records are at opposite ends of the scale.

References

  1. Wu J, Pei F, Zhou L, Li W, Sun R, Li Y, Wang Z, He Z, Zhang X, Jin X, Long Y, Cui W, Wang C, Chen E, Zeng J, Yan J, Lin Q, Zhou F, Huang L, Shang Y, Duan M, Zheng W, Zhu D, Kou Q, Zhang S, Liu Y, Yao C, Shang M, Peng S, Zhou Q, Cheng KK, Guan X, TESTS study collaborator group. The efficacy and safety of thymosin α1 for sepsis (TESTS): multicentre, double blinded, randomised, placebo controlled, phase 3 trial. BMJ. 2025. PMID 39814420 DOI 10.1136/bmj-2024-082583
  2. Wu J, Zhou L, Liu J, Ma G, Kou Q, He Z, Chen J, Ou-Yang B, Chen M, Li Y, Wu X, Gu B, Chen L, Zou Z, Qiang X, Chen Y, Lin A, Zhang G, Guan X. The efficacy of thymosin alpha 1 for severe sepsis (ETASS): a multicenter, single-blind, randomized and controlled trial. Crit Care. 2013. PMID 23327199 DOI 10.1186/cc11932
  3. Dalmasso G, Charrier-Hisamuddin L, Nguyen HT, Yan Y, Sitaraman S, Merlin D. PepT1-mediated tripeptide KPV uptake reduces intestinal inflammation. Gastroenterology. 2008. PMID 18061177 DOI 10.1053/j.gastro.2007.10.026
  4. Kannengiesser K, Maaser C, Heidemann J, Luegering A, Ross M, Brzoska T, Bohm M, Luger TA, Domschke W, Kucharzik T. Melanocortin-derived tripeptide KPV has anti-inflammatory potential in murine models of inflammatory bowel disease. Inflamm Bowel Dis. 2008. PMID 18092346 DOI 10.1002/ibd.20334
  5. Teichman SL, Neale A, Lawrence B, Gagnon C, Castaigne JP, Frohman LA. Prolonged stimulation of growth hormone (GH) and insulin-like growth factor I secretion by CJC-1295, a long-acting analog of GH-releasing hormone, in healthy adults. J Clin Endocrinol Metab. 2006. PMID 16352683 DOI 10.1210/jc.2005-1536
  6. Ionescu M, Frohman LA. Pulsatile secretion of growth hormone (GH) persists during continuous stimulation by CJC-1295, a long-acting GH-releasing hormone analog. J Clin Endocrinol Metab. 2006. PMID 17018654 DOI 10.1210/jc.2006-1702
  7. US National Library of Medicine. PubMed, searched for CJC-1295 and CJC 1295 on 23 August 2026. Both searches return the same set. Three published human studies carry the name, PMID 16352683, PMID 17018654 and PMID 19386527, and all three used the version with the albumin-binding linker. No published human study of the version without that linker was located under any of its names. PubMed. 2026. PubMed search
  8. US Food and Drug Administration. KPV-related bulk drug substances: briefing document for the Pharmacy Compounding Advisory Committee, review dated 12 May 2026. The agency's own literature search across PubMed, Embase, DailyMed, Drugs@FDA and the adverse event reporting system located no human study of KPV. FDA advisory committee materials. 2026. FDA briefing document