Wiki

Ground Truth Wiki

Ground Truth Peptides and the science of grading evidence

How an independent reference grades 42 research peptides on human evidence, safety, real-world signal and placebo, and never collapses them into one number.

What Ground Truth Peptides is and the problem it solves

Ground Truth Peptides is a free, independent, evidence-graded reference for research peptides. As of October 2026 it covers 42 compounds, and it sells nothing: no products, no cart, no affiliate links, no sponsor [2][3]. It went live at groundtruthpeptides.com on June 24, 2026 [15].

The problem it answers is where the information comes from. Almost everything written about peptides comes from someone selling them: a vendor, an affiliate, a clinic with a program to fill [13][15]. The Common Ground record adds that the independent material that existed was thin or tied to vendor discount codes, and that the nearest direct competitor, Reptides, is a paid, login-gated wiki at about seventy dollars a year. Neither grades the evidence, separates scientific support from real-world signal, or says what it would take to verify a claim [15]. The grade is the only product, and it is free [13].

How the grading works

Ground Truth never reduces a compound to one number. One letter would hide the only question that matters, which is proven for what [1]. Every compound answers the same four plain questions, and each keeps its own letter or level with a plain-English read beside it [1][12]:

  1. Does the science back it? The evidence grade.
  2. Do real people feel it? The real-world signal, always labeled as anecdote.
  3. Is it safe? The safety grade.
  4. Could it be placebo? A placebo-risk read on every profile.

A fifth item, legal access, sits beside the four and is deliberately not a grade. It answers whether there is any lawful way to get the compound, and it moves when regulators move [4][12].

The evidence axis, A to E. The evidence grade is the highest level of human evidence a compound has reached. Animal data does not move it [4].

  • A, proven in humans: multiple human randomized controlled trials, or FDA-approved.
  • B, human-trial signal: at least one human randomized controlled trial.
  • C, early human data: human pilot or observational studies only, no randomized trial.
  • D, animal only: promising in animals, no meaningful human evidence.
  • E, theoretical: in vitro, mechanistic, or anecdote, with no controlled data.

The rule that animal data does not raise the grade is the center of the system. A compound can look spectacular in rodents and still grade D [6]. As of October 2026, across the graded index, 6 compounds reach evidence A, 14 reach B, 9 sit at C, 11 at D and 2 at E. Most of the popular recovery and longevity peptides cluster in the low-evidence corner, which the site calls its central finding [3].

The safety axis, A to F. Safety is a composite roll-up, not an absence of bad news. A is well-tolerated with extensive human safety data. B is characterized, with real but manageable risks under monitoring. C is thinly characterized: no red flags observed, but human exposure is small. D is uncharacterized, with little or no human safety data either way. E is flagged for documented serious risk or a major red flag. F is known to cause serious harm [4]. The grade is built from six shown inputs: the volume of human safety data, the severity of known adverse events, theoretical mechanistic risk, the availability of long-term data, the regulatory signal, and the width of the therapeutic window [3].

The real-world signal. For most of these compounds the human trials barely exist, so Ground Truth weights real-world evidence more heavily than a medical journal would, and says what kind and what it is worth. Athletes and bodybuilders count as the de facto trials, first-hand practitioner experience counts, and third-party sellers are referenced by name with the conflict flagged [1]. All of it lands in the "do real people feel it?" lens, labeled as anecdote and weighted up "because the science is thin, never because it beats a trial" [1]. Each profile rates the signal on volume, consistency and source credibility [12].

The confound watch. Almost nobody runs a peptide alone. People run it on testosterone replacement, on a GLP-1 drug, mid-cut, on a real training program, and each of those has large, proven effects of its own. Before crediting a peptide, Ground Truth asks what the known commodities are already doing, and every profile carries a confound watch [1]. The site's long-form writing makes the arithmetic explicit, citing a 1996 testosterone trial [17][20] and the two large GLP-1 trials, STEP 1 [18] and SURMOUNT-1 [19]. In STEP 1, semaglutide produced a mean 14.9 percent body-weight loss at 68 weeks against 2.4 percent on placebo. In SURMOUNT-1, tirzepatide produced up to 20.9 percent at 72 weeks against 3.1 percent [3]. Against effects that size, a peptide stacked beside them is noise until the commodity it rides on is subtracted.

The adversarial pass. Before a grade ships, it goes through a pass whose only job is to break it. That pass hunts for the disconfirming study, the trial that was quietly cancelled or missed its endpoint, the retraction, the replication that never came, and the vendor money behind a "study" being used to sell something. If the grade survives, it ships. If it does not, it moves before anyone sees it [1]. Marketing looks for the one paper that supports the claim. This looks for the one that kills it [1].

Conflict first. The first question asked of any source is not "is it credible?" but "who profits if you believe it?" Where information and profit line up, the site flags it, on the experts it cites, the sellers it references and on itself [1].

A worked example: BPC-157

BPC-157 grades evidence C and safety C, with the one-line read "Promising, anecdote-led" [12]. The reasoning is on the page. There is no completed randomized trial in humans. A Phase I trial (NCT02637284) was registered in 2016 and never reported results [21]. The earliest documented human use is an uncontrolled 2021 case series of 16 knee-pain patients [22]. A 2024 pilot in 12 patients and a 2025 intravenous safety pilot in two healthy adults followed, and the profile counts roughly 30 humans across all published human reports [12]. Each claim on the profile carries its own small grade: the healing claims for tendon, ligament and muscle grade D because the data are overwhelmingly rodent, while the gut-lining claim grades C [12].

It labels the mechanism "assumed," meaning documented in animal cells and tissue but untested at these doses in people [12].

Why the evidence is thin: the economics

Ground Truth's position is that the evidence gap is an economics story and not a safety verdict. Untested is not disproven, and it is not safe [6].

The comparison the site uses is semaglutide against BPC-157. Semaglutide brought in $29.3B in 2024, roughly 70 percent of Novo Nordisk's revenue, is patented, and was studied across dozens of large trials. BPC-157's sequence is in the public domain, no patent worth pursuing exists, and about 30 humans in total have been studied [6]. A single Phase 3 trial costs $11 to $53M depending on the field, 6.7 percent of drugs that enter human testing reach approval, and the path takes about 10.5 years [6]. A trial is only a rational bet if a patent lets the sponsor recoup it. Naturally occurring, long-published sequences are nearly impossible to claim as a new composition of matter, so the trial that would generate the evidence is the trial no for-profit sponsor funds [6]. The peptides that earned approval are owned, branded products. Tesamorelin is a patented analog approved on trials of 816 patients [6].

The same page names who profits from the blank: compounding pharmacies, telehealth platforms, wellness clinics, gray-market vendors, and influencers and affiliates. It also cuts the other way. There is no buried positive trial, and the same economics block dedicated safety trials and adverse-event tracking, so "few reported harms" is weak reassurance, just as "few reported benefits" is weak condemnation [6].

The placebo lens

Placebo is real and neurobiological, running on dopamine, endogenous opioids and endocannabinoids. It moves subjective outcomes such as pain, fatigue, energy and mood far more than objective ones such as labs or tumor size. In one migraine study cited by the site, placebo accounted for at least half of total relief. Open-label placebos, where the person is told the pill is inert, still work, roughly 20 percent better than no treatment, which means "I can feel it working" cannot be the evidence [7]. The site does not treat "could be placebo" as an insult: if a person feels better, that is real to them. The cost is money and certainty, because a person paying peptide prices for an expectation may not learn what actually worked [7][11].

The site's answer is measurement. Its six-step framework is a specific goal and objective metric, one variable at a time, a baseline or control, objective measurement, blinding where possible, and replication, on then off then on again [7].

What can actually be tested

The testable page asks, for all 42 compounds, whether a blood marker reads the drug's action, whether it is only a proxy, or whether no routine lab can tell. The answer: 6 are directly testable, 9 are testable by proxy, and 27 of 42 are not testable by bloodwork [5]. The directly testable group includes semaglutide and tirzepatide, read through HbA1c and fasting glucose, and several growth-hormone secretagogues, read through IGF-1 [5]. For most of the rest the effect is local, subjective or intracellular. A compound can be untestable for efficacy and still testable for harm [5].

Verification and sourcing

Ground Truth applies a five-test verification standard to any gray-market vial: HPLC purity, mass-spectrometry identity, LAL endotoxin, heavy metals and sterility [8]. Mass spectrometry catches substitution, the most dangerous failure, and endotoxin is invisible to HPLC [8]. The site's rule is never to use a vendor-recommended tester, since a lab chosen and paid by the seller is marketing, not verification [8]. Its Certificate of Analysis check has four steps: match the batch number, check the date, confirm every column including net peptide content, and verify the lab independently [8]. It reports that roughly one in three gray-market products fails basic quality testing on identity, purity or quantity [9].

The freshness engine

A grade is a read as of the last review, and new evidence moves it [1]. Daily, the site watches PubMed alerts per compound and new independent-lab batch data. Weekly, it watches scientists, clinicians and communicators, with their conflicts flagged. Monthly, it watches regulatory status. Every item is triaged one of three ways, update a profile, open a new one, or kill it as noise, and logged with its source and a credibility tag, so a change to a grade can be traced to why [10]. The site lists 29 live sources [10].

The rule that follows is that regulation does not move a grade. In July 2026 an FDA advisory committee voted on seven peptides, recommending six for the 503A Bulks List and voting one down. The evidence did not change, so no grade moved, and the legal-access column changed instead [4][12][15]. The vote on BPC-157 was 8 to 6 with one abstention at the July 23 to 24 meeting, FDA's own briefing documents recommended against adding it, and as of October 3, 2026 FDA had proposed no rule, so none of the six can yet be lawfully compounded [10][12].

History and milestones

  • 2023. The FDA restricts 19 wellness peptides to Category 2, pulling them out of routine compounding [11].
  • April 2026. The FDA moves 12 peptides out of Category 2 [11].
  • June 24, 2026. Ground Truth Peptides goes live with 42 graded compounds [15].
  • July 23 and 24, 2026. The advisory committee votes on seven peptides, and not one Ground Truth grade moves [4][12][15].
  • October 3, 2026. Regulatory status last rechecked, with no proposed rule from FDA [12][15].

The site's longer history runs from 1901, when the peptide bond was defined, through insulin in 1922, the first human peptide therapy, to the GLP-1 era. Its required "first documented human use" field is blank for most wellness peptides, which it calls the most honest thing it can show [11]. It puts the scale in perspective: the body runs on an estimated 7,000 peptides, and under 100 are FDA-approved drugs [16].

Who it serves

It serves people deciding whether a peptide claim is worth money, and anyone who needs a citable reference. It is an educational resource, not medical advice: dosing is "reported, not prescribed" [2][13]. Its tools include a reconstitution calculator, a monthly cost planner with no checkout, a research assistant that answers only from the graded and cited profiles, and a tracker for a reader's own lab results [3][14].

The people

Ground Truth was researched, designed and built by Common Ground [12][15]. Jesse Fowler built it [15]. It was built in conjunction with Dr. Matt Zanis, PT, DPT, FAAOMPT, the founder of Rooted in MVMNT, a Duke-trained doctor of physical therapy and fellowship-trained manual therapist who maintains contract advisory with US Olympic organizations [15]. Dr. Zanis needed a place to send his own clients for accurate, current peptide research, separate from anything he or Common Ground sells. The site gives him a page as lead contributor. It does not put him on the homepage and does not call him a peptide authority [15].

Frequently asked questions

What is Ground Truth Peptides? It is a free, independent reference that grades the evidence on 42 research peptides. It sells no products, takes no vendor money and carries no affiliate links [2][3][15].

How does Ground Truth grade peptides? Each compound gets an evidence grade from A to E and a safety grade from A to F, plus a real-world signal and a placebo read, and the four are never combined into one score. The evidence grade is the highest level of human evidence reached [1][4].

Why does a grade not rise when animal studies are strong? The evidence grade counts human evidence only, so strong animal data leaves a compound at D [4].

Does an FDA advisory vote change a grade? No. A vote about who may legally prepare a compound is not evidence that it works. In July 2026 the legal-access column changed and no grade did [4][12].

Can a blood test show whether a peptide is working? For 6 of the 42, yes, and for 9 more by proxy. For 27 of 42, no routine blood test can confirm the effect [5].

Why is there so little human research on peptides? The site says the reason is economic: sequences that cannot be patented rarely attract the tens of millions a human trial costs. Untested is not disproven, and it is not safe [6].

Is Ground Truth medical advice? No. It is an educational resource, and it says to talk to a licensed clinician [2][13].

Sources

  1. Ground Truth Peptides, How we grade, and where we're biased. https://www.groundtruthpeptides.com/method
  2. Ground Truth Peptides, home page. https://www.groundtruthpeptides.com/
  3. Ground Truth Peptides, llms.txt. https://www.groundtruthpeptides.com/llms.txt
  4. Ground Truth Peptides, the grade board. https://www.groundtruthpeptides.com/board
  5. Ground Truth Peptides, What you can actually test. https://www.groundtruthpeptides.com/testable
  6. Ground Truth Peptides, Follow the money. https://www.groundtruthpeptides.com/follow-the-money
  7. Ground Truth Peptides, Is it the peptide, or is it placebo? https://www.groundtruthpeptides.com/is-it-placebo
  8. Ground Truth Peptides, Verification standard. https://www.groundtruthpeptides.com/verification
  9. Ground Truth Peptides, Sourcing and quality. https://www.groundtruthpeptides.com/sourcing
  10. Ground Truth Peptides, What we track, and when. https://www.groundtruthpeptides.com/updates
  11. Ground Truth Peptides, A short history of peptide medicine. https://www.groundtruthpeptides.com/history
  12. Ground Truth Peptides, BPC-157 profile. https://www.groundtruthpeptides.com/peptides/bpc-157
  13. Ground Truth Peptides, What this is. https://www.groundtruthpeptides.com/about
  14. Ground Truth Peptides, Questions. https://www.groundtruthpeptides.com/faq
  15. Common Ground wiki, Ground Truth Peptides: the site that earns its grades. https://wiki.commonground.ventures/work/ground-truth-peptides/
  16. Ground Truth Peptides, How little anyone knows. https://www.groundtruthpeptides.com/the-unknown
  17. Ground Truth Peptides, Writing, where the confounder essay is cited. https://www.groundtruthpeptides.com/writing
  18. STEP 1 semaglutide trial, PMID 33567185. https://pubmed.ncbi.nlm.nih.gov/33567185/
  19. SURMOUNT-1 tirzepatide trial, PMID 35658024. https://pubmed.ncbi.nlm.nih.gov/35658024/
  20. Bhasin, New England Journal of Medicine, 1996, PMID 8637535. https://pubmed.ncbi.nlm.nih.gov/8637535/
  21. ClinicalTrials.gov, NCT02637284, Phase I oral safety and pharmacokinetics trial of BPC-157. https://clinicaltrials.gov/study/NCT02637284
  22. 2021 uncontrolled BPC-157 knee-pain case series, PMID 34324435. https://pubmed.ncbi.nlm.nih.gov/34324435/

Who built this site

The Ground Truth site and this wiki were created by Common Ground. Read how Common Ground built it, and the related work on record.