Peptides from bench to biohack

A dozen amino acids in the right order can tell a cell to build a blood vessel. That is the appeal. What the approved ones actually do, what the research ones have not shown yet, and where the line between them falls.

Last updated 22 August 2026

Peptide basics and evidence literacy

PC-ART-001 Reference 18 August 2026 6 min read

Residues on a backbone. A peptide is 2–50 of these; past roughly 50 the same chemistry is called a protein.

A ball-and-stick molecular model of a short peptide chain lying on a dark slate bench beside a stack of blank index cards. — A model of the backbone. Everything the record can say about a peptide is built on a chain this simple.

The divide that matters first

Start with the thing that makes peptides worth the attention. A chain of a few dozen amino acids walks up to a receptor the body already uses and tells the cell what to do next. No brute force, no foreign machinery — the message arrives in a language the cell has been speaking since before you were born. That is a genuinely elegant piece of biology, and it is why this field moves as fast as it does.

Now the boundary, which is messier than anyone admits. Two definitions of "short" are in circulation and they disagree. The textbook convention is 2 to 50 residues. US regulation draws its line at 40 — above that a molecule is a protein and takes a different approval route entirely. So between 41 and 50 residues you have a compound that is a peptide by convention and a protein by law.

Either definition covers a drug with millions of patient-years behind it and a vial sold with "not for human consumption" printed on the side. The chemistry does not tell those two apart. The evidence does, and that is the line this entire site is built along.

A small clear glass vial with a crimped aluminium cap, holding a shallow layer of white lyophilised powder, standing on a brushed steel bench. — A vial of lyophilised powder. Beautiful, inert, and completely silent about which side of that line its contents fall on.

Approved therapeutics

Carry phase 3 trial data, an approved label, and manufacture under cGMP. Semaglutide and tirzepatide are the widely prescribed examples.

Indications and dosing are on a public label
Identity and purity are a regulatory requirement
Prescription-only, with a traceable supply chain

Research compounds

Sold explicitly as not for human consumption. BPC-157 and CJC-1295 sit here. Much of the preclinical work is real; what is absent is the human record.

No approved label, so no established dose
Purity varies between sellers and between batches
The uncertainty is as much supply chain as pharmacology

Three ways a peptide acts

This is the part worth getting excited about. A peptide does not force a system open. It knocks on a receptor the body already uses, and everything downstream is the body's own cascade — which is how a molecule this small moves something as large as appetite or blood supply. Three patterns cover most of what is on file.

Receptor agonism
GLP-1 analogues bind receptors in the gut, pancreas and hypothalamus at once, slowing gastric emptying and altering appetite signalling. Attaching a fatty-acid chain lets the molecule ride serum albumin. The semaglutide label names that binding as the principal mechanism behind its half-life of about a week.
Hormone secretagogues
Rather than supplying growth hormone from outside, these act on GHRH or ghrelin pathways so the pituitary releases its own pulse. CJC-1295 with ipamorelin is the pairing most often described — and the corpus holds CJC-1295 as two separate records, with and without the DAC modification. The DAC version's half-life was measured at 5.8 to 8.1 days in healthy adults. The version without it has no comparable human figure on file.
Tissue repair and angiogenesis
Repair-associated peptides upregulate vascular endothelial growth factor and modulate inflammatory cytokines, which in animal models speeds cell migration and matrix remodelling.

The peptide matrix

Five families, and the sweep of them is why people fall down this particular rabbit hole: appetite, growth, repair, cognition, mitochondria. Read the last column strictly, though. What a row is studied for describes the approved compounds in it, and says nothing about the research-grade ones sitting beside them.

Families on file

5 categories · 13 compounds

Category	Key compounds	Mechanism / target	Studied for
Metabolic	Semaglutide, Tirzepatide, Retatrutide	GLP-1 / GIP / glucagon receptor agonism	Weight reduction and glycaemic control, in phase 3 trials
Growth secretagogues	CJC-1295 DAC, CJC-1295 no-DAC, Ipamorelin, Sermorelin	Pituitary GHRH / ghrelin receptor activation	Endogenous GH and IGF-1 pulse, body composition
Healing and recovery	BPC-157, TB-500	Fibroblast activation, actin binding, angiogenesis	Tendon and ligament repair — animal models
Cognitive	Semax, Selank	ACTH fragments, BDNF modulation, nasal CNS access	Focus and anxiolysis, largely outside Western trials
Mitochondrial and longevity	SS-31, Epitalon	Cardiolipin stabilisation, telomerase	Cellular energy and oxidative stress

Where the evidence comes from

Here is the tool that makes everything above readable. Every record on this site is graded on one ladder, strongest evidence at the top, and the rubric is published so you can argue with any grade we hand out. A rung is a statement about the best study on file — not a verdict on whether a compound works. Plenty of genuinely promising things sit low down because nobody has run the trial yet.

Two analysts in white coats standing at a laboratory bench, reading chromatography traces on a monitor. — A chromatogram is a purity claim about one batch on one afternoon. It is not evidence that the compound does anything at all.

Evidence ladder

Strongest to most exploratory

01
Human RCT
Randomised controlled trial in humans.
54 compounds
02
Human observational
Cohort, case-control, cross-sectional or registry data in humans.
17 compounds
03
Human case series
Clinician-documented individual human cases, uncontrolled.
4 compounds
04
Animal in vivo
Whole-animal studies.
57 compounds
05
In vitro
Cell culture, tissue, or cell-free systems.
16 compounds
06
Theoretical
Mechanistic reasoning, with no direct measurement for this claim.
1 compound
07
Computational
In silico only — docking, structure prediction, QSAR, machine learning.
3 compounds

A rung is the strongest study on file, not a verdict. 7 of the 7 rungs are the top of some compound's record, and 37 of 189 records carry no tier at all — that is the finding, not a gap in ours. The rubric is published, so any grade here can be audited.

Handling conventions

The practical layer, and it is more interesting than it sounds: most of what goes wrong with a peptide happens between the vial and the syringe rather than in the pharmacology. These are the conventions the field works to and what the diluent record supports. They describe practice. They are not instructions, and nothing here recommends that anyone use anything.

A grey tray on a pale surface holding a capped vial, a sealed alcohol swab and an insulin syringe, arranged in a row. — Three objects, one clock. The 28-day count starts at first puncture — not at reconstitution, and not at delivery.

Reconstitution

Lyophilised powder is mixed with a diluent, usually bacteriostatic water, before it can be drawn. Mass divided by diluent gives concentration; dose divided by concentration gives volume.

Beyond-use date

A multi-dose vial is dated at first puncture and discarded 28 days later, unless the manufacturer states another date. Bacteriostatic water qualifies because its benzyl alcohol is a preservative; sterile water and plain saline carry none and get no such window. The 28 days is a rule about the vial, not a storage temperature.

Certificates of analysis

A COA states one batch's test results, issued by whoever ran them — usually the manufacturer's own quality unit, not an independent lab. It covers that batch and nothing else.

Baseline measurement

Metabolic panel, fasting glucose and insulin, lipids, thyroid and IGF-1 are the markers usually recorded before and during use, because they are what any change would show up in.

Glossary

Every term used above, defined the way this site uses it. No word should be doing load-bearing work in an article you cannot check.

Acylation: Attaching a fatty-acid chain so the molecule binds serum albumin, which slows clearance and extends half-life.
Angiogenesis: Formation of new blood vessels.
Beyond-use date: The date a multi-dose vial is discarded after first puncture — 28 days by convention, unless the manufacturer states another.
Certificate of analysis: One batch's test results, issued by whoever ran them — usually the manufacturer, not an independent lab.
Endogenous: Produced by the body itself.
Exogenous: Introduced from outside the body.
Growth hormone secretagogue: A compound that prompts the pituitary to release its own growth hormone rather than supplying it.
Half-life: The time for the concentration in the body to fall by half.
Lyophilised: Freeze-dried to powder, the form peptides are stored in before reconstitution.
Reconstitution: Mixing lyophilised powder with a diluent to make a solution.
Residue: A single amino acid unit within a chain.
Stacking: Combining two or more peptides in one protocol.

Sources and evidence

14 sources

Kaprive JF, Krishnamurthy K. Biochemistry, Peptide. StatPearls, updated 28 August 2023Reference text