Compounded peptides

Compounded peptides: what they do and what the evidence shows

Compounding pharmacies prepare custom medicines from raw ingredients for individual prescriptions. A few peptides can already be compounded legally. Most of the popular ones, such as BPC-157 and TB-500, cannot: in 2023 the FDA flagged them for significant safety risks, and in July 2026 its advisory committee voted to let six back in, against FDA staff advice. That vote is not binding. Tap any peptide for its effects, evidence and risks. For approved peptide drugs and compounds in trials as well, browse the peptide library.

How to read the evidence scale

LabAnimalsSmall human studiesLarge trialsFDA approved

Filled steps show the strongest kind of evidence published for each peptide.

Peptide Evidence Index

How sure can we be?

Each peptide scored from 0 to 10 on two questions: how solid is the biology behind it, and how certain is it that it helps people. Peptide Analysis assessment of the published evidence, October 2026.

Ranking

Index = 25% biology + 75% evidence in people. Tap a peptide for the full analysis.

1Melanotan IIReceptor biology well mapped; small controlled trials show tanning and erections, and a close relative is FDA approved. Real safety concerns.Biology 8In people 5.56.1 2SermorelinWell-defined hormone fragment; reliably raises GH and IGF-1 in trials and was FDA approved. Benefit for normal ageing unproven.Biology 9In people 56.0 3NAD+Central to cell metabolism; oral precursors raise NAD+ in trials with small metabolic signals. IV use barely studied.Biology 8In people 45.0 4LL-37Human antimicrobial peptide studied by many independent groups; one positive placebo-controlled wound trial.Biology 7.5In people 44.9 5GlutathioneCore cellular antioxidant; oral doses raise body stores, but trials show little clinical benefit.Biology 8.5In people 34.4 6GHK-CuConsistent lab and skin data, much from one group; small cosmetic cream studies. No trials of injections.Biology 6In people 3.54.1 7MOTS-cStrong mouse papers from independent journals; only a small four-week trial of a modified analogue in people.Biology 7In people 2.53.6 8SemaxRodent brain-growth-factor data; approved in Russia on small local stroke studies of uncertain quality.Biology 4.5In people 33.4 9BPC-157Hundreds of animal studies, mostly from one lab; human data are uncontrolled case series.Biology 6.5In people 23.1 10TB-500The parent protein is well studied, but the fragment sold online has no human trials.Biology 5.5In people 12.1 11Emideltide (DSIP)Natural role and gene never established; 1980s human sleep trials were small and mixed.Biology 2In people 22.0 12KPVClear gut mechanism in a few mouse colitis studies; no human trials found.Biology 5In people 12.0 13EpitalonTelomere and lifespan claims come from one research network with little independent replication.Biology 3In people 1.51.9 14DihexaRodent memory data from one group; a related drug failed in Alzheimer's. Never tested in people.Biology 4In people 0.51.4 15PEG-MGFEarly muscle-growth findings not reproduced by an independent lab; no human trials.Biology 2In people 0.50.9

How the scores work

Biological evidence rates how well the mechanism is understood and how strong the laboratory and animal work is: quality of the studies, whether independent groups have reproduced them, and whether the substance is a well-characterised molecule in the human body. Certainty it helps people rates human data for the uses each peptide is marketed for: 0 means no human studies, 2 to 3 uncontrolled reports or weak trials, 4 to 5 at least one positive placebo-controlled trial or a reliable effect on a biomarker, 6 to 7 consistent controlled trials, and 8 to 10 the level of evidence behind approved medicines. Scores reflect evidence, not safety: a peptide can work and still carry real risks, as Melanotan II does.

Why 75% for human evidence. Only studies in people can show whether a peptide actually helps, and promising animal results often fail to carry over: most drugs that enter human trials are never approved. Biology keeps 25% of the weight because human data for most of these peptides are sparse, and well-reproduced laboratory work makes small human reports more credible. Ties are broken by the evidence in people.

The Reddit Index

What Reddit users report, and what the evidence shows

All 15 compounded peptides, ranked by how effective Reddit users report them to be, next to our own ranking based on the published evidence.

Reddit rankPeptideConfidence in the Reddit signalPeptide Analysis rank

Reddit ranking based on user reports in Reddit peptide communities, October 2026. The confidence rating reflects how many reports exist and how consistent they are. Peptide Analysis rank: our ranking from 1 to 15 by the Peptide Analysis Index (25% biological evidence and 75% evidence in people, scored 0 to 10). Reddit is not a clinical trial: posts cannot rule out placebo, natural recovery or a mislabelled product, and satisfied users post more often. Not medical advice.

Already allowed in compounding

Can be prescribed and compounded today

Sermorelin

Can be compounded

Growth hormone-releasing hormone analogue

Once an FDA-approved medicine for children; now compounded for adults.

How sure can we be?

Ranking

What Reddit users report, and what the evidence shows

Already allowed in compounding

Sermorelin

Glutathione

NAD+

Voted on in July 2026

BPC-157

TB-500

KPV

MOTS-c

Semax

Epitalon

Emideltide (DSIP)

Up for review by February 2027

GHK-Cu

LL-37

Dihexa

Melanotan II

PEG-MGF

A warning on "research peptides"