Every peptide page on the internet contains numbers, and most of them are missing something: a unit, a frequency, a route, or, most often, a source. This page is the key to reading them. It explains what the units mean and how they trip people up, why a syringe unit is not a dose, how a per-kilogram trial figure becomes a flat one, and how to grade a number by where it came from. It is the method behind Peptifact's dosage-reference pages, which apply it compound by compound — from the one compound in the catalogue that now has an approved label to a molecule whose entire indexed literature is four preclinical papers. It recommends nothing. How brands appear on this site is set out on our disclosure page.
The five parts of a dose figure
A complete dose statement answers five questions: how much, by what route, how often, for how long, and according to whom. "250 mcg" answers one. "250 mcg subcutaneously twice daily for six weeks, per a clinic's protocol page dated 2025-10-30" answers all five, and the last answer is the one that tells you how much the other four are worth. Peptifact's dosage pages put every figure in a table with those five columns; a figure that cannot fill them is reported as incomplete rather than completed by guesswork.
Units: mass, volume, and the trap between them
Mass. Micrograms (mcg, µg) and milligrams (mg): 1,000 mcg is 1 mg. Most research-peptide figures are in micrograms; TB-500 and GHK-Cu pages use milligrams; approved GLP-1 schedules are in milligrams. The two abbreviations differ by one letter and by a factor of a thousand, and pages that mix them without care produce thousandfold errors. FDA has reported hospitalizations from dosing errors with compounded GLP-1s in which patients "measuring and self-administering incorrect doses" was the mechanism; the unit trap is part of it.
Volume, disguised as a dose. A U-100 insulin syringe is graduated in units, where 100 units is 1 mL and one unit is 0.01 mL. That is a volume scale. The mass in one unit is the solution's concentration divided by 100: 25 mcg per unit at 2.5 mg/mL, 50 mcg at 5 mg/mL, 250 mcg at 25 mg/mL. So "10 units" is 250 mcg, 500 mcg or 2.5 mg depending only on the vial and the diluent volume, and a figure stated in units without the concentration behind it is not a dose at all. Forum posts state units constantly; our reconstitution page has the tables that turn them back into mass, given the concentration.
Per kilogram versus flat. Trials often dose by body weight; clinic pages state flat figures. To compare them, multiply: the CJC-1295 trial's "30 or 60 microg/kg" is 2.25 to 4.5 mg for a 75 kg adult; the ipamorelin surgical trial's 0.03 mg/kg is 2.25 mg; the pediatric sermorelin label's 30 mcg/kg is 600 mcg for a 20 kg child. The conversion is arithmetic, and it usually shows that the trial figure and the clinic figure are not the same number, the same form, the same route or the same population. Our CJC-1295/ipamorelin page works the examples.
Sometimes the conversion cannot be done at all, and that is worth recognising rather than forcing. Where a trial states a rate — so much per kilogram per hour, delivered by pump — the figure needs a body weight and a duration before it becomes a quantity, and the kisspeptin record is the clearest case on this site of a compound whose entire published human dose set is expressed in units a vial label cannot carry. A per-kilogram figure from an animal study is the commonest case: it does not become a human quantity by multiplying by body weight, which is why our FOXO4-DRI record reports its only dose — 5 mg/kg in mice — without converting it.
Per dose, per day, per week. "250 mcg twice daily" is 500 mcg a day and 3.5 mg a week; "2.5 mg twice weekly" is 5 mg a week; "2.4 mg once weekly" is 2.4 mg a week. Pages that state a per-dose figure beside a per-week figure for another compound invite a comparison that means nothing. The dosage pages state the frequency the source states and do not convert it.
A figure describing how long the compound stays in the blood is graded the same way, and most of the ones in circulation do not survive it: the per-compound half-life record separates the label figures from the mouse, dog and test-tube figures printed beside them.
Grading the source
The number's value is set by how it was found. Peptifact uses six tiers, and each figure on the dosage pages carries its tier:
| Tier | Kind of source | What stands behind the number | Examples on this site |
|---|---|---|---|
| 1 | FDA-approved label | Dose-finding trials, an approved indication, validated reconstitution and storage | Wegovy, Zepbound, EGRIFTA; the withdrawn pediatric Geref label |
| 2 | Trial protocol or peer-reviewed study | Tested in humans for a stated purpose, not approved; may be a different form, route or population from the use in question | Retatrutide 4/9/12 mg (TRIUMPH); CJC-1295 with DAC 30–60 mcg/kg; intravenous ipamorelin; intra-articular BPC-157 |
| 3 | Compounding pharmacy dispensing information | Practice under a prescriber's direction; a product that exists at a stated strength | Empower and Olympia sermorelin pages |
| 4 | Clinic protocol page | One practice's convention, usually acknowledged as such | Beverly Hills Rejuvenation Center; Tucson Wellness MD |
| 5 | Vendor or protocol site | Marketing or community convention; the better ones say so | Peptide Dosing Protocols; FormBlends |
| 6 | Forum post | Anonymous, unsourced, often stated in syringe units | Not quoted on this site |
Every dose record on this site is now indexed by that grade in one place: the dosage chart sorts all 29 compounds and blends into the tier of their best available source, and shows how few of them sit anywhere near the top. It also shows the limit of the scale — the hCG label is a tier-1 source that describes its own regimens as ones "advocated by various authorities" rather than as established figures.
The tiers are about provenance, not truth: a tier-1 label can be for a different indication than the one a reader has in mind, and a tier-5 page can accurately report what a tier-2 trial used. The point of the grade is that the reader can see which is which. The sorting question is simple: does a human dose-finding study exist for this compound, by this route, for this use? For semaglutide, tirzepatide and tesamorelin, yes. For retatrutide, a phase 2 dose-ranging trial exists and approval does not. For BPC-157, TB-500, GHK-Cu by injection, CJC-1295 without DAC, subcutaneous ipamorelin and adult sermorelin, no, and every figure for them is a convention, however many pages repeat it.
The question can also fail one step earlier than that. FDA's July 2026 review of semax could find no pharmacokinetic study in humans by any route, and it has identified no human exposure data of any kind for dihexa: for a compound in that position, a stated figure is not a weak source to be graded but no human measurement of the substance at all. Our page on the cognitive peptides sets out what the record holds for those two and for selank, noopept, cerebrolysin and P21.
The agreement trap
For compounds without a trial, the clinic, vendor and protocol pages agree with one another to within a factor of two: BPC-157 at 250 to 500 mcg, TB-500 at 2 to 2.5 mg, GHK-Cu at 1 to 2 mg, sermorelin at 200 to 300 mcg. It is tempting to read that consistency as consensus. It is copying. The reverse failure is just as quiet: the two figures quoted for DSIP are both written "25", one in nanomoles per kilogram and one in micrograms, and they differ by about eighteen percent. A convention that originated in a forum post in the 2010s and has been reproduced by every guide since is still one data point, and the guides that are honest about it, several are, say "community-derived" or "no universally accepted human dosage" in their own text. The dosage pages quote that language when a source uses it. The clearest single case of the trap is duration: the cycle lengths every guide states disagree with one another, appear in no approved label, and are absent from the titles of all 141 registered trials of these compounds.
What "commonly used" means on this site
Occasionally there is no figure to report at all. Searching the human literature for a published epitalon dose returns studies of a bovine pineal extract rather than the synthetic peptide, and the right answer is to say so rather than to quote the extract's numbers. It means the figure that appeared most often on the pages we opened on a stated date, with those pages cited and linked. It does not mean validated, safe, effective, prescribed by us or recommended by us. The distinction is the entire basis on which these pages exist: they report what sources state, graded by source, so that a reader can decide what weight to give a number, and they leave what to do with it to the reader and their clinician.
The dosage-reference pages
- BPC-157: one human case series, a thousandfold animal range, and the 250 to 500 mcg convention.
- TB-500: a fragment with no human study; the full protein's trials at very different doses.
- CJC-1295 and ipamorelin: trial doses by weight and by vein; clinic doses by convention.
- Tesamorelin and sermorelin: a current label, a withdrawn label, and the adult convention that grew after it.
- GHK-Cu: topical trials without stated concentrations; injectable figures without trials.
- MOTS-c: no label, no completed trial, and a registration whose sponsor filed seven others.
- PT-141: an approved label with a real dose, for a population and a purpose the same label excludes.
- Semax: hundreds of papers, no registered trial, and doses published as percentages rather than milligrams.
- Selank: a search count that halves twice when you restrict it, leaving two Russian-language randomised papers in one journal.
- Melanotan II: one registration in the world, and an approved relative whose label converts to nothing.
- AOD-9604: more published papers on detecting the compound in urine than primary studies of what it does in people.
- Semaglutide: four brands, two routes, five labelled ladders, and a manufacturer's own statement that two of its oral products are not interchangeable by milligram.
- Tirzepatide: two labels listing identical strengths, and three different ceilings depending on indication and age.
- Retatrutide: no label anywhere; every circulating figure is an arm label from one 338-participant trial.
- Cagrilintide: 43 registrations from one sponsor, and the only phase 3 testing it alone is in children.
- Kisspeptin: a full published human dose record, every figure a per-kilogram infusion rate that no vial label can express.
- VIP / aviptadil: an NIH trial dosing in picomoles per kilogram per hour, which converts to a whole course weighing under a milligram.
- Glutathione: a ten-to-fifteen-minute intravenous half-life, and an FDA advisory committee that overruled its own agency 8 to 5.
- Noopept: one figure, 20 mg a day, from a single unblinded Russian study, with no registered trial under any of its three names.
- Tesofensine: four oral doses, twenty years, three sponsors, thirteen registered studies and no phase 3.
- How peptides are reconstituted: the procedure per an FDA label and the arithmetic for any vial.
- Are peptides legal?: what the "research use only" label on the vial these figures describe actually requires, quoted from the regulation.
Sources and dates
The figures cited as examples on this page are drawn from the dosage-reference pages linked above, each of which lists its sources and the date each was opened (2026-09-05). Corrections go to the contact page.