Peptides in 2026: the biohacker's no-BS field guide

Key takeaways

• Tirzepatide hit roughly 20.9 percent mean weight loss at 72 weeks in SURMOUNT-1 . Semaglutide 2.4 mg hit roughly 14.9 percent at 68 weeks in STEP-1 . Retatrutide reached roughly 24.2 percent at 48 weeks in phase 2 . No other peptide bucket has trial numbers in that weight class.
• BPC-157 has one published 16-patient knee-pain retrospective and zero RCTs; the rodent mechanism work in tendon, ligament, and GI injury models is real . TB-500 as a label covers two different molecules (full-length thymosin beta-4 around 4963 Da, versus the 17-23 LKKTETQ fragment around 889 Da); the CAS number and molecular weight on the COA are the only way to know which is in the vial .
• Rapamycin reliably extends mouse median and maximum lifespan by roughly 10 to 25 percent across NIA Interventions Testing Program cohorts . Mannick's everolimus trial improved influenza-vaccine antibody response in adults 65 and older . No human mortality outcome trial has reported.
• Compounded GLP-1s are not the labeled drug. FDA has flagged dosing errors with compounded semaglutide as a specific patient-harm signal separate from the safety profile of Wegovy or Zepbound .
• The four failure modes that actually put people in clinics are identity mismatch, lot-to-lot purity drift, endotoxin contamination, and reconstitution dose drift. None are solved by reading more PubMed; all are solved by the supply chain, the math, and the log.

Verdicts by bucket: the at-a-glance

The reader-payoff version, before the analysis. Each compound below is named in the body sections with its trial citations and its caveats; this section is the scorecard you came for.

• Fat loss, tirzepatide (Zepbound, Mounjaro). Phase-3 strong: roughly 20.9 percent mean weight loss at 72 weeks in SURMOUNT-1, around 91 percent of treated participants reached at least 5 percent loss . FDA-approved for obesity.
• Fat loss, semaglutide 2.4 mg (Wegovy). Phase-3 strong: roughly 14.9 percent mean loss at 68 weeks in STEP-1, around 86 percent reached at least 5 percent loss . FDA-approved for obesity.
• Fat loss, retatrutide. Best pipeline signal: roughly 24.2 percent at 48 weeks in phase 2 with no clear plateau at the top dose . Not yet approved.
• Fat loss, orforglipron. Oral non-peptide GLP-1, late phase. If approved on the sponsor's timeline, the access conversation changes more than the efficacy conversation.
• Fat loss, AOD-9604. Failed human obesity trials at the doses tested . Do not shop it as a GLP-1 substitute.
• Fat loss, tesamorelin (Egrifta). FDA-approved only for HIV-associated visceral fat accumulation . Off-label fat-loss positioning is not what the trial population looked like.
• Healing, BPC-157. Real preclinical signal across tendon, ligament, GI, and angiogenesis pathways. Human chart: one 16-patient knee-pain retrospective and a small number of unblinded case reports. Zero published RCTs .
• Healing, TB-500 / thymosin beta-4. Label covers two different molecules; full-length thymosin beta-4 has a published phase 1 and a small post-MI pilot, the LKKTETQ 17-23 fragment is the rarer case on current COAs. Verify CAS and molecular weight before assuming which one is in the vial .
• Healing, GHK-Cu. Skin and wound-healing mechanism. Sits on the FDA 503A interim bulks list with route-of-administration concerns flagged for the injectable; PCAC consult signaled before end of February 2027 .
• Longevity, rapamycin (sirolimus). 10 to 25 percent mouse lifespan extension across NIA ITP cohorts . Mannick low-dose mTOR inhibitor improved vaccine response in adults 65+ . Human longevity use is entirely off-label.
• Longevity, NMN and NR. Oral dosing raises blood NAD+ in human trials . No translation to a lifespan or healthspan outcome yet.
• Longevity, MOTS-c. Mitochondrial-derived 16-residue peptide identified in 2015; mouse data on insulin sensitivity and metabolic homeostasis . No registered exogenous human outcome trial.
• Longevity, senolytics (dasatinib + quercetin, fisetin). Feasibility and target-engagement pilots in idiopathic pulmonary fibrosis and diabetic kidney disease. No outcome benefit yet.

The honest 2026 landscape, in one paragraph

If you tried to read the peptide market in 2026 from social media, you would conclude that everything works, the regulators are gatekeepers, and the only question is which influencer to trust. If you tried to read it from FDA's homepage, you would conclude that everything is unapproved, dangerous, and basically a scam. Both readings are wrong, and they are wrong the same way: they ignore that the evidence is wildly uneven across the three buckets people actually shop in. Fat loss has phase-3-grade evidence on a small group of injectable GLP-1s. Healing has a real preclinical signal on a couple of peptides with almost no human trial coverage. Longevity has decades of mechanism work and mouse lifespan data, and the human outcome trials are mostly still running. Treating those three buckets as one conversation is the first mistake.

What follows is the no-BS roundup: where each bucket sits in 2026, which compounds carry the evidence, and which tools turn a peptide protocol from a coin flip into a calculated risk. The inline links point at the deeper per-compound and per-topic articles in the catalog when you want to drill in.

Healing: BPC-157, TB-500, and the marketing-vs-evidence gap

The healing peptides are the loudest part of the consumer market and the thinnest part of the human trial base. BPC-157 is a 15-amino-acid sequence derived from a gastric protein. Rodent pharmacology across tendon, ligament, muscle, and GI injury models shows faster repair, with mechanism work pointing at angiogenesis pathways including VEGFR2 and nitric oxide signaling. A 2025 narrative review concluded the preclinical record is real and that BPC-157 remains investigational because the human work has not been done at scale . The human record is one 16-patient retrospective for intra-articular knee pain and a small number of unblinded case reports. No published randomized placebo-controlled trial exists at the time of writing.

TB-500 is the messier story, but not in the direction the older write-ups claim. The name has historically been used as a catch-all for two different molecules: full-length thymosin beta-4 (43 amino acids, the molecule with the published phase 1 healthy-adult trial and the small post-myocardial-infarction pilot) and a short N-acetylated LKKTETQ fragment covering positions 17-23 of thymosin beta-4. Older equine doping-control work characterized the marketed veterinary TB-500 product as the N-acetylated LKKTETQ fragment , and that finding got generalized into a 'TB-500 sold to consumers is the fragment' narrative. The 2026 reality on certificates of analysis is the other way around: most product currently sold under the TB-500 label is the full 43-residue thymosin beta-4, and the short 17-23 fragment is the less common case. The two are easy to tell apart on a COA by CAS number and molecular weight (full-length thymosin beta-4 is roughly 4963 Da; the LKKTETQ 17-23 fragment is roughly 889 Da). The takeaway is not that the vial is automatically the wrong molecule. The takeaway is that 'TB-500' as a label does not pin down which version you have, so identity verification on the COA is the only way to know which one you are injecting. The BPC-157 evidence map and the Wolverine stack write-up walk the per-compound evidence in detail; the per-compound library entries are BPC-157 and TB-500.

Sport status is also worth flagging. WADA places BPC-157 under S0 non-approved substances and thymosin beta-4 with its derivatives (TB-500 named as an example) under S2 peptide hormones, growth factors, and related substances, both prohibited at all times in tested sport . If you compete under any tested-sport body, this category is a non-starter regardless of evidence.

Fat loss: the GLP-1 era and the adjacent compounds nobody should confuse for them

Fat loss is where the science actually shows up in 2026. The injectable GLP-1 class is the only part of this conversation with phase-3-grade randomized human trial coverage. Semaglutide 2.4 mg weekly (Wegovy) produced roughly 14.9 percent mean weight loss at 68 weeks in STEP-1, with around 86 percent of treated participants losing at least 5 percent . Tirzepatide (Mounjaro for diabetes, Zepbound for obesity) is a dual GIP and GLP-1 receptor agonist and produced roughly 20.9 percent mean weight loss at 72 weeks at the top dose in SURMOUNT-1, with around 91 percent reaching 5 percent loss and 57 percent reaching 20 percent loss .

Retatrutide is the triple agonist (GLP-1, GIP, and glucagon). In an Eli Lilly phase 2 trial run by Jastreboff and colleagues, the top dose produced roughly 24.2 percent mean weight loss at 48 weeks without a clear plateau, suggesting longer dosing or higher doses may produce additional loss . Retatrutide is not approved yet. Cagrilintide plus semaglutide (CagriSema) and high-dose Wegovy sit in the same phase-3 readout window. Orforglipron is the oral non-peptide GLP-1 receptor agonist that, if approved on the timeline the sponsor has been signaling, will change the access conversation more than the efficacy conversation.

What does not belong in the fat-loss conversation with GLP-1s: AOD-9604, a 16-residue fragment of the C-terminus of human growth hormone that was developed as an oral anti-obesity peptide and failed to beat placebo in human trials at the doses tested . Tesamorelin (Egrifta) is a real drug, FDA-approved specifically for HIV-associated visceral fat accumulation, and the Stanley 2014 JAMA trial showed durable visceral adipose tissue reduction in that population . It is not a general anti-obesity drug, and the off-label fat-loss positioning the consumer market gives it is not what the trial population looked like.

The tirzepatide vs semaglutide head-to-head, the retatrutide triple-agonist piece, and the orforglipron oral-GLP-1 article have the per-compound trial tables. The GLP-1 reconstitution and titration risk article has the dose-math primer.

Longevity: rapamycin, NAD precursors, MOTS-c, senolytics

Longevity is the bucket where mechanism is strong, rodent data is strong, and human outcome data is mostly still in pilot phase. Rapamycin (sirolimus) is FDA-approved for transplant rejection (Rapamune) and for tuberous-sclerosis-associated angiofibromas (Hyftor). The longevity conversation is entirely off-label. The NIA Interventions Testing Program reported rapamycin reliably extends median and maximum mouse lifespan by roughly 10 to 25 percent across multiple cohorts and starting ages . The Mannick rapalog immune-aging trials showed a low-dose mTOR inhibitor (everolimus) improved influenza-vaccine antibody response in adults 65 and older . The PEARL trial is the first dedicated longevity-focused human pilot. None of these are mortality outcome trials.

NAD precursors (NMN and NR) are the supplement-shelf version of the longevity story. The mechanism is real: NAD+ declines with age, NAD+ is the substrate for sirtuins and PARPs and the mitochondrial electron transport chain, and oral NR and NMN both raise blood NAD+ in human trials . What the trials do not show, yet, is a translation from raised NAD+ to a lifespan or healthspan endpoint in humans. The bet is that the rodent and mechanism data carry. That bet is reasonable. It is not finished.

MOTS-c is a 16-amino-acid mitochondrial-derived peptide encoded in the mitochondrial 12S rRNA region. Lee and colleagues identified it in 2015 and showed it regulates insulin sensitivity and metabolic homeostasis in mice . Human plasma MOTS-c rises with exercise. No registered human trial has tested exogenous MOTS-c against a hard endpoint. Senolytics (dasatinib plus quercetin, fisetin, others) are the most promising new mechanism and the most overhyped in the consumer market. Early human pilots in idiopathic pulmonary fibrosis and diabetic kidney disease have shown feasibility and target engagement, not yet outcome benefit. Treat the longevity bucket as informed bets on mechanism, not as finished medicine. The rapamycin off-label longevity piece, the NAD precursors evidence snapshot, the MOTS-c mitochondrial-derived peptide article, and the senolytics early human trials article walk the per-compound state.

The four risk vectors that decide whether a protocol survives

Strip out the regulatory abstractions and the safety problem is concrete. Four risk vectors keep showing up when peptide products move from research supply chains into consumer use. None of them are about the molecule. All of them are about the vial.

1. Identity mismatch. The label says BPC-157, TB-500, GHK-Cu, or CJC-1295. The vial may or may not contain that molecule, and even when the family is right, the specific form can vary. 'TB-500' as a label has been used for both full-length thymosin beta-4 (roughly 4963 Da) and the shorter 17-23 LKKTETQ fragment (roughly 889 Da), so a CAS number and molecular weight on the certificate of analysis are what actually confirm identity .
2. Purity variability lot to lot. Even when identity is correct, the percentage of intact peptide, the truncated and oxidized side products, and the residual synthesis solvents vary by manufacturer and by batch. Research-grade synthesis is not held to USP pharmacy compounding standards or to FDA-inspected drug-manufacturing standards. There is no public audit trail you can read.
3. Endotoxin and bioburden. Injectable products require endotoxin testing because gram-negative bacterial lipopolysaccharide can cause fever and worse at very low microgram doses, even when sterility is achieved. USP Chapter 797 sets endotoxin limits for pharmacy-compounded sterile preparations. Research-chemical peptides do not run that test as a routine release criterion. Unexplained fever shortly after injection should put endotoxin on the differential.
4. Dose drift from reconstitution. A lyophilized 5 mg vial reconstituted with the wrong volume of bacteriostatic water gives the wrong dose. Stability after reconstitution depends on temperature, light, and time. Most reconstitution guides on social media are screenshots of someone else's screenshot. The dose you think you took is not necessarily the dose your tissue received. FDA has flagged compounded-GLP-1 dosing errors as a specific harm signal .

The research-use-only sticker is the most misunderstood phrase in the peptide market. It is not an FDA safety category. It is a seller-applied label that signals the product is sold for in-vitro research, not for human use, and exists to limit the seller's liability under federal drug law. It does not put the product under FDA quality systems. Compounded drugs, including peptide compounds, are not FDA-approved and have not been evaluated by FDA for safety, effectiveness, or quality before sale . Research-chemical products sit one layer further from oversight than that. The peptide hype vs safety article walks the full label-meaning conversation; this section is the short version.

FDA continues to flag specific unapproved products, including peptide-class ones, in consumer-facing communications when the harm signal is loud enough. The agency's tanning-pills page warns that oral canthaxanthin tanning pills are not FDA-approved for tanning and have been tied to adverse effects such as canthaxanthin retinopathy . The May 14, 2026 update to the 503A interim bulks list still treats GHK-Cu by route of administration and flags the injectable route specifically, with FDA signaling it intends to consult PCAC on potential 503A inclusion before the end of February 2027 .

Tools that make a peptide protocol survivable

If the molecule is the easy part and the vial is the hard part, the tools that matter are the ones that make the vial side auditable. None of these are perfect. None are FDA-supervised. Together they are the difference between a calculated risk and a complete leap.

1. Reconstitution math, written down. The number of units on the syringe is not the number of milligrams in the dose. A peptide reconstitution calculator that takes the vial size, the diluent volume, and the target dose, and returns the syringe units, removes the screenshot-of-a-screenshot failure mode. The pepSmart vial tracker and reconstitution calculator exist for this exact reason.
2. A dose log, not a memory. Date, time, vial lot, site, dose, and any acute symptoms within 4 hours. When something goes sideways at week 6, the log is the only artifact that can tell you whether the issue tracks with the molecule, the lot, or the site. A symptom diary on a notes app does most of this job; a structured tracker does it better.
3. Baseline and follow-up bloodwork on a fixed cadence. For GLP-1s: a metabolic panel, an HbA1c, a lipid panel, and a lipase before titration and after dose changes. For rapamycin: a CBC, a metabolic panel, a fasting lipid panel, and a fasting glucose every 8 to 12 weeks. For testosterone-adjacent or HPG-axis-active compounds: total testosterone, free testosterone, SHBG, LH, FSH, estradiol, and prolactin on a structured cadence. The pepSmart bloodwork doc has the per-class panels.
4. Cold-chain shipping and a refrigerator, not a windowsill. Lyophilized peptides are stable longer than reconstituted ones, but neither survives a hot mailbox. Storage and stability are not pharmacy-class problems; they are user-class problems, and a fridge with a thermometer beats opinions.
5. A clinician you tell the truth to. Withholding the peptide history from a provider when something goes wrong does not protect you. It just makes the chart less useful. Persistent fever, sudden visual changes after a melanocortin product, new or changing pigmented lesions, priapism, signs of pancreatitis, or anything systemic and unexpected belongs in a clinic, not a forum thread.

Editorial summary

The honest read on the 2026 peptide market is not the corporate cover-your-ass version (every research peptide is poison, defer to FDA in all things) and it is not the influencer version (regulators are gatekeepers and the testimonial is the trial). It sits in between, and that is uncomfortable for anyone who wants a clean rule.

Fat loss has a real trial base on a small group of GLP-1s, and the consumer drug-market problem is compounded-product quality, not whether the molecule works. Healing peptides have a preclinical signal worth taking seriously and a human chart too small to justify finished-medicine claims. Longevity peptides and rapamycin-class drugs have strong mechanism, strong rodent data, and pilot-grade human data, which makes them informed bets, not prescriptions.

The bigger near-term safety story for the consumer is not the molecule. It is the vial. Identity, purity, endotoxin, and dose drift are the failure modes that put people in clinics, and they are solved by the supply chain, the math, and the log, not by reading more PubMed. For research and educational purposes only. Not medical advice.

Related tools

• Tirzepatide dose calculator - Run tirzepatide-focused vial draw math.
• GLP-1 conversion calculator - Convert a GLP-1 mg dose to U-100 units and ml.
• GLP-1 ramp planner - Preview a linear educational dose-step table.
• Peptide half-life calculator - Estimate single-dose decay from cited half-life constants.
• PK simulator overview - Public overview of the Pro pharmacokinetic simulator.
• Protocol builder overview - Public overview of the Pro protocol builder.