Under 1% is not a formulation failure — it is the expected result of six barriers multiplied together. The exceptions are five specific ways of cheating one barrier at a time.
The gastrointestinal tract is, by design, a machine for dismantling peptides into amino acids — that is how dietary protein becomes nutrition. An oral peptide drug has to run that machine backwards: survive chemistry, cross a gel, cross a sealed cell sheet, dodge export pumps, and pass the liver, in sequence. The key insight is that the failures multiply. Bioavailability is the product of the fractions surviving each step, so six steps survived at a respectable 30 to 50 percent each still lands under one percent.
The field's anchor reviews state the resulting arithmetic directly: the oral bioavailability of whole peptides and proteins is typically less than 1 percent, and in some cases lower than 0.1 percent (Wang et al. 2026, PMID 41585434; Baral and Choi 2025, PMID 40284395). Less than one percent is not a bad formulation. It is the default outcome — and every exception in this guide works by defeating one specific step, or by not needing absorption at all.
Fasted stomach fluid runs at pH 1.2 to 3.0 — acid unfolds peptides and pepsin starts cutting, preferentially beside aromatic residues. In simulated gastric fluid, a benchmark study found insulin about 34 percent intact at 10 minutes, about 1 percent at 30 minutes, and undetectable at 60 (as reported in PMID 41585434 — a bench model, not a human stomach measurement). Surviving that earns entry to the small intestine, where near-neutral pH is exactly where trypsin, chymotrypsin, elastase and brush-border peptidases work best. Escaping the acid trades one hostile environment for another.
A continuously shed gel coats the gut wall. Peptides are trapped in it by charge, hydrogen bonding and size, and the loose outer layer is constantly cleared — a molecule can be perfectly intact and still never reach a cell (PMID 41585434).
Between cells, tight junctions seal the gaps — the openings are measured in ångströms, the molecules in nanometres, and junctions cover under 1 percent of the surface. Through cells, hydrophilic peptides partition poorly into lipid membranes, and what is taken up can be chewed by intracellular enzymes on the way through (PMID 41585434). Both routes are, for a typical peptide, nearly closed.
Enterocytes run export pumps (P-glycoprotein is the canonical one) that return absorbed molecules to the lumen. Whatever survives absorption drains to the portal vein and through the liver before reaching the body — where gut-wall and hepatic metabolism take a further cut, measurable and clinically important even for the field's best-case molecule, cyclosporine (PMID 1958443).
Rybelsus proves a peptide pill can work systemically — at roughly 1 percent efficiency, under strict conditions, with an exceptionally potent, long-half-life molecule. It is the exception that validates the arithmetic.
The mechanism is local and gastric. The tablet co-formulates semaglutide with 300 mg of an enhancer, SNAC (the amount selected because exposure peaked there in first-in-human testing, PMID 30565096). The definitive mechanism study (Buckley et al. 2018, Science Translational Medicine, PMID 30429357) showed absorption happens in the stomach, not the intestine, confined to the area touching the tablet; SNAC buffers the pH in that tiny zone (locally suppressing pepsin) and the peptide crosses through the stomach cells — transcellular — with no effect on tight junctions. Popular accounts that say SNAC "opens the junctions" have the mechanism backwards; junction-opening is the other approved enhancer's trick, two sections down.
The number. Population pharmacokinetic modelling puts bioavailability at about 0.8 percent under the labelled dosing conditions (PMID 33969456); the FDA label range is 0.4 to 1 percent. Within-person variability in bioavailability was 137 percent, damped at steady state only because daily dosing and a one-week half-life average it out. The daily 3/7/14 mg doses exist because roughly 99 percent of each tablet never reaches the blood — the pill is mostly a delivery tax. Efficacy at that efficiency was proven in PIONEER 1 (703 patients, 26 weeks, PMID 31186300): placebo-adjusted HbA1c fell 0.6, 0.9 and 1.1 percent at 3, 7 and 14 mg, with 2.3 kg of weight loss at the top dose. The FDA approved it in September 2019 — the first oral GLP-1.
Why the ritual is so strict. Take it fasted, with no more than 120 mL of water, then nothing by mouth for 30 minutes — because each clause maps to a measured effect. In the food-effect study, the fed group showed limited or no measurable semaglutide exposure at all, while every fasting subject absorbed (PMID 34080123); longer post-dose fasting produced significantly more exposure, so 30 minutes is a floor, not an optimum. SNAC's protection is a concentration-dependent phenomenon at the tablet's contact patch: food, fluid, or a second tablet dilutes or displaces it, the protected zone collapses, and the peptide is digested like lunch.
Build your protocol, log every dose, monitor your body's response, and get reminders so you never miss a dose.
Start Tracking FreeCyclosporine: structure does the work. This 11-residue fungal peptide needs no enhancer, because its architecture defeats proteases and membranes simultaneously: it is cyclic, and extensive backbone N-methylation strips away the exposed amide groups that proteases grip and that prevent membrane crossing (reviewed in PMID 27178381). In a head-to-head experimental comparison, N-methylated cyclosporine achieved 22 percent oral bioavailability where simple cyclic leucine peptides managed 4 to 17 percent (PMID 25313329) — the modifications, not the ring alone, do the lifting. In children studied before transplant, mean oral availability was 20.6 percent, ranging from 10.8 to 34.1 percent, with gut-wall metabolism contributing to its lowness and variability (PMID 1958443). Wider ranges quoted in secondary sources (up to 70 percent) could not be verified against a primary source for this guide and are not printed. Even the success story loses most of its dose and needs blood-level monitoring.
Desmopressin: 0.1% is enough if the drug is potent enough. Desmopressin's name encodes two edits to vasopressin — deamination at position 1 and a D-arginine at position 8 — that buy protease resistance: in human intestinal contents, native vasopressin was rapidly destroyed while 50 percent of desmopressin remained intact at 35 minutes (PMID 8532599). Site-by-site measurement in six volunteers found bioavailability of 0.19 percent from the stomach, 0.24 percent from the duodenum and 0.19 percent from the jejunum, falling to 0.03–0.04 percent further down (PMID 8359185); the marketed tablet's label figure, 0.08 to 0.16 percent, sits inside that data. Those studies dosed 400 µg into the gut to do the work of 4 µg intravenously — a hundredfold tax, rational only because vasopressin receptors respond at picogram-per-millilitre levels. And one elegant experiment quantifies the barrier itself: adding the protease inhibitor aprotinin raised desmopressin's duodenal bioavailability fivefold, from 0.09 to 0.46 percent (PMID 8858277). The enzymes are the wall; the experiment measures the wall.
If the receptor faces the gut lumen, the entire absorption problem disappears. Linaclotide (14 amino acids) activates guanylate cyclase-C on the intestinal surface, driving fluid secretion and transit; it is minimally absorbed by design, dosed at 145 to 290 µg daily for constipation indications, with diarrhoea as its signature adverse effect (PMID 24917937). Plecanatide, a 16-residue analogue of the natural peptide uroguanylin, went further: in its single-dose Phase 1 study (72 subjects, doses up to 48.6 mg), there was no measurable systemic absorption at any dose, with an assay sensitive to 1 ng/mL (PMID 23625291), and it won its first approval in January 2017 (PMID 28255961).
The lesson cuts both ways. For gut-local targets, non-absorption is a safety feature — systemic side effects are structurally excluded. And the trick is simply unavailable for any target outside the gut wall, which is most of the targets people care about. "Oral" and "absorbed" are different claims, and a peptide can be a successful oral drug by failing the second one completely.
Dosing cheat sheet, reconstitution reference, and cycle planning — delivered to your inbox.
Mycapssa (approved June 26, 2020) embodies the second enhancer philosophy: where SNAC goes through stomach cells, its Transient Permeability Enhancer — an oily suspension built on sodium caprylate — briefly opens the between-cells route in the intestine. In rats, the formulation transiently rearranged the tight-junction protein ZO-1 and let paracellular markers up to 70 kDa pass, reversibly and dose-dependently; nine months of daily dosing in monkeys produced only minor findings (PMID 24558008).
The pivotal trial (CHIASMA OPTIMAL, PMID 32882036) switched 56 acromegaly patients previously controlled on injectable therapy to oral octreotide or placebo for 36 weeks: IGF-1 control was maintained in 58.2 percent on oral therapy versus 19.4 percent on placebo (p=0.008), and 75 percent of the oral group completed the trial on capsules. Note the shape of the indication the FDA granted — long-term maintenance in patients already responding to and tolerating the injectable. It is a switch-and-maintain product, not a starting therapy: the direct regulatory shadow of modest, variable absorption. No verified oral-octreotide bioavailability percentage was located for this guide, so none is printed — the defensible statement is that absorption is low, variable, and managed by titrating dose against IGF-1 response.
The marketing syllogism goes: "BPC-157 is stable in gastric juice, therefore oral BPC-157 is absorbed, therefore capsules work like injections." Each arrow is a separate claim with a separate evidence status, and they deserve to be separated.
What is published. BPC-157 is a synthetic 15-residue peptide (sequence GEPPPGKPADDAGLV) derived from a gastric-juice protein. Its developers describe it as "stable in human gastric juice" — in narrative reviews written by the originating laboratory itself (Sikiric et al. 2012, PMID 22300085), never independently replicated, with no duration figure in the sources this guide verified. The group does dose it orally in rodent GI-injury models and reports effects; those experiments measure injury outcomes, not blood levels of intact peptide.
What is not published — and this list is the finding. The only formal pharmacokinetic study, by an independent group, dosed rats and dogs intravenously and intramuscularly only: half-life under 30 minutes, IM bioavailability 14–19 percent in rats and 45–51 percent in dogs, rapid breakdown to fragments — and no oral arm (He et al. 2022, PMID 36588717). There is no human oral pharmacokinetic study at any dose in any formulation; no animal oral bioavailability measurement; no validated oral-to-injectable dose conversion; and the most recent critical review counts fewer than 30 human subjects across three uncontrolled pilot studies in the entire clinical record, with no completed Phase II and no standardised preparation (PMID 42198317).
The three distinctions. Stability is not absorption: desmopressin is also comparatively protease-resistant and still yields about 0.1 to 0.2 percent. Absorption is not efficacy at a distant tissue: an effect inside the gut after oral dosing — the original ulcer and colitis context — would not require systemic absorption at all, as linaclotide proves in reverse. And absence of measurement is not proof of zero: the honest sentence is that BPC-157's developers report gastric stability and rodent oral activity, and no study has ever published how much intact BPC-157 reaches the blood after oral dosing, in any species. Anyone selling certainty beyond that sentence is selling past the evidence in one direction or the other.
Keep this section's tense discipline: everything beyond the five exceptions above is developmental, and most of it is preclinical.
The realistic near-term future is not a universal peptide pill. It is more exceptions, each engineered for one molecule and one barrier at a time — which is exactly what the five exceptions already are.
Nothing above tells you where to buy, because sourcing is the part that varies most. Whoever you use, require the same three things:
The vendor we recommend — full panel published per batch. A reader discount use it at checkout.
For research use only. Not for human consumption. 21+.
Build your protocol, log every dose, monitor your body's response, and get reminders so you never miss a dose.
Start Tracking FreeA few can — five mechanistically distinct exceptions have produced approved products: enhancer-based absorption (oral semaglutide, oral octreotide), intrinsically hardened molecules (cyclosporine, desmopressin), and gut-local agents that never need absorbing (linaclotide, plecanatide). For an unmodified peptide with a target outside the gut, the expected oral bioavailability is under 1%, sometimes under 0.1% (PMID 41585434). The exceptions are engineered, molecule by molecule; there is no general peptide pill.
Because its enhancer, SNAC, works only in a tiny buffered zone where the tablet touches the stomach wall, and only at high local concentration. In the food-effect study, fed subjects showed limited or no measurable semaglutide exposure while every fasting subject absorbed the drug (PMID 34080123), and longer post-dose fasting produced significantly more exposure. Food or extra fluid dilutes and displaces the enhancer, the protected zone collapses, and gastric enzymes digest the peptide before it can cross.
Nobody knows, because it has never been measured. The only formal pharmacokinetic study dosed BPC-157 intravenously and intramuscularly in animals — no oral arm exists (PMID 36588717) — and no human oral pharmacokinetic study has been published at any dose. The gastric-stability claim traces to narrative reviews by the developing laboratory (PMID 22300085), and stability is not absorption: the comparatively protease-resistant desmopressin still reaches the blood at only about 0.1 to 0.2% (PMID 8359185). Capsule doses in the market are not derived from any published absorption data.
Per milligram, dramatically: oral bioavailability is about 0.8% under ideal labelled conditions (PMID 33969456), which is why the pill is a daily 3 to 14 mg dose against a once-weekly subcutaneous dose roughly an order of magnitude smaller. At its studied doses it is nonetheless effective — PIONEER 1 showed placebo-adjusted HbA1c reductions up to 1.1% at 14 mg in 703 patients (PMID 31186300). The honest comparison is not strong versus weak; it is a ~99% delivery tax paid for needle-free dosing, with strict fasting rules as part of the price.
For most peptides, not on any visible horizon. The barriers multiply, and the entire history of the field is five narrow exceptions rather than a general solution: two enhancer products with Phase 3 data, two structurally exceptional molecules, and gut-local drugs that succeed by not being absorbed. Ingestible injection devices are the most conceptually promising route around the problem and remain animal-stage (PMIDs 30733413, 31591601). Expect more exceptions, engineered one molecule at a time — not a universal fix.
Disclaimer: This guide is for educational and informational purposes only and is not intended as medical advice, diagnosis, or treatment. The compounds discussed are not FDA approved for human use. Always consult a qualified healthcare provider before starting any new supplement or peptide protocol. StackTrax does not sell peptides or supplements directly — purchase links go to third-party vendors. StackTrax is not responsible for the products, quality, or business practices of any third-party vendor.
© 2026 StackTrax, LLC. All rights reserved.
Privacy · Terms · Do Not Sell or Share My Personal Information
StackTrax guides cover peptides and compounds that are not FDA-approved for the uses discussed. The dosing, reconstitution, and safety information is compiled from published research and community protocols for educational purposes only.
Before using any compound mentioned here, consult a qualified healthcare provider. StackTrax does not sell, prescribe, or recommend these substances for personal use.