Six is not a dodge. Each rule counts something real and each produces a different number, so the only wrong answer is a single figure given without saying what it counted.
Thousands, if the question is about nature, and the first section below gives that figure with its source. The number is real and almost never the one people want. What follows it is always how many you can buy, how many are medicines, or how many have anything behind them.
Aren't there thousands of peptides?
Yes, and the count depends on which thousands. The Antimicrobial Peptide Database held 6,309 peptides as of January 1, 2026, split into 3,379 natural, 2,290 synthetic and 373 predicted, with the natural entries drawn from six kingdoms of life.1 Antimicrobial activity is one function among many, so that figure describes a slice.
No clean total exists for peptides in nature, and the reason is structural. Every protein broken down in digestion yields peptides, so the set has no boundary at the bottom end, and the number of possible short sequences is combinatorial: twenty amino acids arranged ten at a time give more than ten trillion possibilities before anything biological is asked. Counting what exists produces a definition dressed as a number.
Neither of the two therapeutic-peptide reviews read for this page offers a figure for naturally occurring peptides, despite both surveying the field.23 They count drugs instead. This page makes the same move for the same reason: the countable sets are the ones somebody has drawn a boundary around.
How many compounds does this library track?
This library tracks 383 compounds, as of the October 2026 capture.4 No count on this page is wider, or looser, because the rule behind it is permissive: anything sold and discussed as a peptide earns an entry, whether or not it survives closer inspection.
A library built that way has to carry things that fail the definition, because somebody looking up a name needs to find it. The alternative is a reference that goes quiet exactly when a reader needs it most.
An entry is more than a name in a list. Each carries a route, a half-life where one has been measured, what the compound is used for, what the human evidence consists of, and a status saying how it is available. Compounds that cannot support those fields still get an entry, with the gaps showing, because an honest blank is more use than an omission. Several of the 383 exist on this site mainly to tell a reader that almost nothing is known about them. Saying so is a legitimate job for a reference and a strange thing to count as coverage.
Why are 36 of them not peptides at all?
Because they are sold beside peptides and asked about in the same breath. 36 of the 383 are labeled as non-peptides, leaving 347 that are.4
The 36 are not a random assortment. Most are small molecules, the kind that work by mouth precisely because they are not chains of amino acids. MK-677 is the clearest case: it raises growth hormone, it is sold beside growth hormone peptides, and it is a synthetic small molecule that survives digestion for exactly that reason. The rest run to a coenzyme in NAD+, a nucleotide in NMN, a synthetic dye in methylene blue, a bile acid derivative, a sulfated sugar chain and a modified amino acid. Every one of them is sold by somebody as part of a peptide protocol.
The pattern behind the group matters more than the examples. A compound tends to end up here when it produces an effect people associate with peptides, through chemistry that has nothing to do with peptides. Nobody is confused about whether a dye is a peptide. They are asking whether it belongs in the stack, and the category label follows the marketing and not the molecule.
Dropping them would make the number look tidier and the reference worse. Somebody searching a name they saw in a stack needs to land on a page that tells them what the thing is, and "not a peptide" is the most useful sentence that page can carry. The plain-language guide sets out that editorial position in full.
Does counting names give the same answer as counting compounds?
No. The gap between the two is wide enough to break a naive count. One compound carries several names, and several compounds can hide behind one name.
Most entries in the library run under at least two names: a code from development, a generic name, sometimes a brand, and often a shortened form the market invented. Tirzepatide and Mounjaro are one compound. TB-500 and thymosin beta-4 are sold as if they were one compound and are not quite the same thing. A count of distinct names in circulation would run far above 383 without a single extra molecule existing.
The reverse case is worse for a counting exercise. Some bare names do not resolve to one compound at all, and CJC-1295 is the standing example, because the version with DAC and the version without behave so differently that treating them as one entry would mislead anyone reading a dose. The library keeps them apart and refuses to resolve the bare term to either. Keeping them apart is right for a reader and awkward for a tidy total.
How many are approved medicines?
76 are approved as drugs in the United States.5 A further 24 are approved somewhere outside the US and not here, and 11 of the US approvals are for products no longer sold.5
Approval is the strictest rule on this page and the one that changes the picture most. Roughly a fifth of the library has been through a regulatory review ending in a marketing authorization. The other four fifths have not, and the gap between those two groups is larger than any difference inside either one.
Approved elsewhere is its own category and gets misused constantly. A compound approved in one country and not another has cleared a real regulator, so the phrase is not empty, and it also does not mean a US buyer is getting that approved product. The 24 in that group are counted separately here for exactly that reason, since folding them into the 76 would let a seller's marketing copy borrow an authorization it does not have.
Withdrawn approvals complicate it further. 11 US approvals cover products no longer marketed, and a compound can be simultaneously approved on paper, absent from pharmacies and available from a research seller. Counting that compound as approved is defensible and so is counting it as unavailable. Both figures appear here for that reason, in place of a net.
How many are still being developed?
Still in development: 47 sit at Phase 3 or later, with 46 in Phase 2 and 16 in Phase 1.5 5 are in Phase 4, meaning post-approval study.5
Development status answers a question approval cannot. A compound in Phase 3 has cleared enough to justify a large trial. The signal is real and it is not a verdict. Nothing about that status predicts approval, and the library contains compounds that sat at Phase 2 for a decade.
Set the 47 beside the 76 already approved in the US and the shape of the field appears. Compounds in late-stage trials run to roughly two thirds of the number ever approved here. By the standards of most drug classes that is crowded. Most of those 47 will not arrive, and the ones that do will take years.
Phase is also a status attached to a program and not to a molecule. One compound can sit at Phase 3 for an indication nobody on this site is interested in while having no program at all for the use people are buying it for. The phase number is real, and it may be answering a different question from the reader's.
How many have been tested in humans at all?
242 have been tested in humans to the extent of carrying a registered trial. Of those, 157 have a trial that posted results.5
Hold those two numbers side by side. Registration means somebody declared an intention to run a study. Posted results mean the study finished and reported. The 85-compound gap between them is the space where trials were registered and then went quiet. Funding, failure and plain abandonment all land there, so a registry entry on its own establishes very little.
Registration is also self-declared and costs nothing to file. Records can be created, never recruit a participant and sit in the registry indefinitely. Sellers quoting "in clinical trials" as evidence are making a claim the registry does not support on its own. Whether a trial finished and said anything is the useful question, not whether one exists.
Tested in humans is doing a lot of work in that heading, too. A registered trial might have enrolled twelve people for a week, and it counts here identically to a three-year study in two thousand. So this count records whether research happened at all. It says nothing about how much, and a reader treating 242 as a depth measure is reading the wrong thing.
How many have nothing behind them?
153 have no published clinical trial, and 45 of those have nothing indexed at all.5
The second figure is the one to remember. 45 compounds in a library of 383 are sold, discussed and dosed with no indexed human research of any kind attached to the name. They are not obscure within the market, and several are among the better-known things on the research-peptide circuit.
Nothing indexed is a stronger statement than no evidence of benefit. It means a search of the trial registries and the published literature returns nothing for that compound in humans, so there is no result to argue about, no sample size to criticize and no protocol to read. Whatever is known comes from animal work, from chemistry, or from what people tell each other.
A reader can hold that against a compound or not. What the number does is separate two groups that marketing copy blends together: compounds whose human trials disappointed, and compounds whose human trials have never happened. Those two groups deserve different treatment from a buyer, and in most marketing copy they get the same silence.
How many can you actually buy?
You can actually buy 175 of the 383, each carrying at least one price listing from a tracked seller, drawn from 14,541 listings across 208 sellers in the capture of October 24, 2026.6
Commercial availability turns out to be a sharper filter than most people expect. More than half the library is not currently for sale anywhere the price index reaches. That group takes in approved drugs no pharmacy lists at a public price, and compounds existing only inside trials.
The listing total points the other way and is easy to misread. 14,541 listings sounds like a vast market, and it describes 175 compounds, so the average compound carries dozens of listings from dozens of sellers. Depth is not breadth. What looks like enormous choice is mostly the same short catalog sold repeatedly. Price comparison works on this market for that reason, since the products are substitutable because they are frequently identical. What reads as a huge market is a narrow one sold many times over, and the catalog is the wider of the two.
How do blends get counted?
Blends get counted as products in their own right. Somebody decided that, and it is not a fact about them. A blend arrives in one vial, dissolves as one solution and gets injected as one dose, so from the buyer's side it is one thing. From the chemistry side it is two or three compounds that happen to share a container.
Counting it both ways would double-count. Counting it only as its components would hide a product people buy by name and search by name. Counting it only as a blend would hide the fact that somebody taking it is taking two or three compounds at doses chosen by whoever filled the vial.
The library lists blends as entries and records their components, so the compound total and the component total differ by design. Any count that does not say which convention it used cannot be checked by the reader.
Which of these numbers is the right one?
The right number is whichever one matches the question being asked, and that is why the page gives six instead of one.
Somebody asking what to take is asking about the 76, or more precisely about the subset carrying an indication that matches their situation. Somebody asking what the research-peptide market consists of is asking about the 175. Somebody asking how much of this is real is asking about the 153, or the 45 inside it. A single headline figure would answer none of those while sounding like it answered all three.
That failure mode is specific and common. A seller writes that there are hundreds of peptides and thousands of studies, both true under some counting rule, and the reader takes it as a statement about the compound in front of them. Neither figure says anything about that one product. Aggregate counts describe a field, and nothing a field-level number says transfers to an individual vial.
Picking a number is therefore the same act as picking a question. Somebody who cannot say which of the six they want has not yet worked out what they are trying to find out, and noticing that is useful before buying anything.
Why does the number keep rising?
The number keeps rising because the library tracks a market. Markets add names faster than they retire them. The inclusion rule is explicit and it is a counting rule like any other on this page: a compound earns an entry when two or more tracked sellers list it. One seller is watched and not built, on the grounds that a single store experimenting with a name is not yet a market.
That threshold decides the total as much as any fact about chemistry does. Lower it to one seller and the library grows immediately, filled with names that may vanish by the next sweep. Raise it to five and the count falls, and the reference stops answering searches for the compound somebody just bought. Two is a judgment about when a name has become real enough to need a page.
A rising number should therefore not be read as more peptides existing. The molecules were already there, and most have been described in the literature for years. What changes is how many of them somebody has decided to sell, and how many sellers it took before the name crossed the line.
Does a bigger library mean better coverage?
Not by itself. The ratio matters more than the total. A library that grew to 500 entries by adding compounds with nothing indexed behind them would be a larger reference and a worse one.
Watch the 45 with no indexed research against the library total. If that group grows faster than the whole, the catalog is drifting toward cataloging marketing instead of medicine, and the growth is measuring how much novelty the market generates rather than how much anyone knows.
One defensible argument runs the other way. Compounds with no research behind them are exactly the ones a buyer most needs a page for, because nothing else they find will say so plainly, and a reference covering only well-studied compounds would go silent where silence does the most damage. Both things hold at once, and the ratio earns attention precisely because it sits between them.
Coverage has a second meaning the total also hides. Any library can hold 383 entries and still be thin on any one of them, so the better measure of a reference is how much each page can say before it runs out of sourced material.
What makes these counts hard to compare with anyone else's?
Comparing counts is hard because every one rests on a rule somebody chose, and most published figures do not state theirs. Two peer-reviewed reviews make the point better than an argument can. A 2019 review in the International Journal of Molecular Sciences reports that "Over 60 peptides have been approved and administrated globally in clinics" and "more than 400 peptide drugs that are under global clinical developments."2 A 2025 review in the same journal reports "over 80 peptide drugs have gained global approval" as of 2023.3
Approvals rising from 60 to 80 across six years is ordinary. The development figures are not. The 2025 review says "more than 200 peptides in clinical development" in its introduction and "650+ candidates in development" in its conclusions, in one paper, without reconciling them.3 Neither number is flagged as wrong, and a reader quoting either one would be quoting the same source.
Counts of approved peptide drugs also move depending on whether insulin analogs are in, whether withdrawn products stay counted, whether approvals outside the US count, and where the peptide-protein boundary sits. The figures on this page state their rule in the same sentence as the number. That makes them comparable across time on this site and hard to set beside a number from somewhere else. That trade was made deliberately, on the view that a number a reader can check beats a number a reader can compare.
What would make the count stable?
A stable count needs a fixed definition. Nobody has agreed on one. The 50-amino-acid line between a peptide and a protein is a convention and not a standard, as the plain-language guide sets out, and the commercial boundary is looser still.
Several bookkeeping questions would have to be settled before any two counts could be compared directly. Somebody has to put the peptide-protein boundary somewhere, and where it goes puts insulin on one side or the other. The plain-language guide gives insulin at 51 amino acids, one residue past the usual convention. Under the rule most often quoted, the most widely used peptide drug in the world is a protein. Each compound has to count once or once per name, settling whether a drug with a code name, a generic name and two brands is one entry or four. Each blend has to count as itself, as its components, or as both.
None of those is a scientific question. Each is a bookkeeping decision that somebody has to make and state, and the reason published counts disagree is that most make them silently. Until that changes, any single figure is a claim about a counting rule. Each rule on this page is stated beside its number so a reader can pick the one answering the question they arrived with.
Common questions
How many peptides are there?
The library tracks 383 compounds as of October 2026, of which 347 are peptides and 36 are not. Different counting rules give different answers, from 76 approved medicines down to 45 with no indexed research at all.
How many peptides are FDA-approved?
76 compounds in this library are approved as drugs in the United States, and 11 of those approvals cover products no longer sold. A further 24 are approved outside the US and not here. The three figures stay separate because folding them together lets a seller borrow an authorization that does not apply where the buyer lives.
How many peptides can you actually buy?
175 of the 383 have at least one listing in the price index, taken from 14,541 listings across 208 sellers in the October 24, 2026 capture. More than half the library is not currently for sale anywhere the index reaches.
How many peptides have no research behind them?
153 have no published clinical trial, and 45 of those have nothing indexed at all. Those 45 are not obscure, and several are well known on the research-peptide market.
Why does the library include things that are not peptides?
36 entries are small molecules, coenzymes, nucleotides and similar compounds sold alongside peptides and asked about in the same breath. Each is labeled as a non-peptide, because a reader searching the name needs a page that tells them what the thing actually is. Leaving them out would make the library go quiet exactly where somebody needs it.
How many peptides are in clinical trials?
47 sit at Phase 3 or later, 46 at Phase 2 and 16 at Phase 1, with 5 in post-approval Phase 4 study. A further count, 242, have at least one registered human trial of any phase.
What is the difference between a registered trial and a published one?
Registration records an intention to run a study. Publication records what the study found. 242 compounds have a registered trial and 157 have one with posted results, so the 85-compound gap covers studies registered and never reported.
Does the number of peptides grow every year?
This library grows because the market adds names, not because new molecules appear. Most of these compounds were described in the literature years before anyone sold them, so the count measures commercial activity as much as chemistry. A name enters when two tracked sellers list it, so the threshold is part of the number.
How many peptide drugs are approved worldwide?
Published reviews put it above 60 in 2019 and above 80 as of 2023, both counting global approvals. This library counts 76 approved in the US specifically. That is a different rule, so the figures are not interchangeable.
How many types of peptides are there?
That depends on the grouping, like every other count here. This library sorts compounds by what people use them for, running from weight loss and metabolic through healing and recovery, longevity, brain and sleep, skin, sexual and hormonal, growth hormone, muscle and immune support. Grouping by chemistry, by receptor or by route would each produce a different set of types and a different number of them.
How many peptides exist in nature?
No usable total exists, because every digested protein yields peptides and the possible short sequences run into the trillions by simple combinatorics. Bounded counts come from curated databases: the Antimicrobial Peptide Database held 6,309 entries at the start of 2026, covering one functional class.
Is there an official list of peptides?
No registry defines the category for commercial purposes. The 50-amino-acid convention separating peptides from proteins is a working rule and not a standard. Two sources counting peptides can disagree without either being wrong.
Why do published counts of peptide drugs disagree so much?
Because each one applies a different rule and most do not state it. Whether insulin analogs count, whether approvals outside the US count and where the peptide-protein boundary sits each move the total by a wide margin.
How many of the approved ones are still sold?
11 of the 76 US approvals cover products no longer marketed. The library keeps them because the name still appears in research and in sellers' marketing copy.
Which number should a buyer care about?
The 175 with listings, narrowed by whatever evidence standard that buyer applies. Availability and evidence are separate questions, and a compound being purchasable says nothing about which of the other counts it belongs to.
How current are these numbers?
The library figures are from the October 2026 capture and the price figures from the sweep of October 24, 2026. Both registries are regenerated on a schedule, and the counts on this page are computed from them instead of typed in.
Sources
- The Antimicrobial Peptide Database (APD), University of Nebraska Medical Center, aps.unmc.edu. Read at source. States that "the APD6 contained 6309 peptides" as of January 1, 2026, comprising 3,379 natural, 2,290 synthetic and 373 predicted entries, with natural entries covering six kingdoms of life. A curated database of one functional class, not a census of peptides. ↩1
- A Comprehensive Review on Current Advances in Peptide Drug Development and Design. International Journal of Molecular Sciences, 2019, mdpi.com. Read at source. "Over 60 peptides have been approved and administrated globally in clinics" and "there are currently more than 400 peptide drugs that are under global clinical developments." ↩1 ↩2
- Therapeutic Peptides: Recent Advances in Discovery, Synthesis, and Clinical Translation. International Journal of Molecular Sciences, 2025, mdpi.com. Read at source. "As of 2023, over 80 peptide drugs have gained global approval" with "more than 200 peptides in clinical development" in the introduction, against "650+ candidates in development" in the conclusions. The two development figures appear in one paper and are not reconciled. ↩1 ↩2 ↩3
- Peptide Decoding compound library, capture of October 2026. Computed from the site's own compound registry, the same file that generates the library listing and
/llms.txt. Entry total, the non-peptide count and the category labels. ↩1 ↩2 - Peptide Decoding evidence index, /evidence. Computed from the site's own trial registry. Approval counts for the US and elsewhere, withdrawn approvals, development phase counts, registered-trial and posted-result counts, and the counts for compounds with no published trial and nothing indexed. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6
- Peptide Decoding vendor price index, /prices, sweep of October 24, 2026. Compounds carrying at least one listing, total listings and seller count. ↩1