The compound library, graded by evidence
Fifty-seven compounds, from approved GLP-1 receptor agonists to peptides with almost no human data. Every one carries a grade, and the grade is the point.
The compound library holds 57 peptides. Each entry covers the mechanism, the human and animal research that exists, dosing ranges reported in the literature, common stacks and contraindications. Every compound carries an evidence grade. When the honest answer is that almost no human data exists, the page says so instead of padding.
What is on each compound page
| Section | What it covers |
|---|---|
| Mechanism | What the compound is and what it is understood to act on |
| Research | The human studies and the animal studies, kept apart |
| Reported dosing ranges | Ranges as reported in the literature, not a recommendation |
| Common stacks | Compounds it is often combined with |
| Contraindications | Who should not use it, and known cautions |
| Evidence grade | How strong the human evidence is, at a glance |
How it works
- Find the compound. Browse or search the 57 entries. The same set is on the website at /compounds, for example semaglutide, tirzepatide or BPC-157.
- Read the grade first. It tells you whether you are reading about a drug with large randomised trials or a compound studied mostly in rodents.
- Add it to a protocol. Any compound in the library can be logged and scheduled in the protocol planner.
Why grade the evidence at all
The word "peptide" covers a very wide range. At one end are drugs such as semaglutide and tirzepatide, approved by the FDA after phase 3 trials with thousands of participants. At the other end are compounds whose evidence is a handful of animal studies, cell experiments and anecdote. Read in the same format, without a grade, those two kinds of page look equally authoritative. That is the problem a grade solves.
Medicine has a standard way of thinking about this. The GRADE approach, set out in the BMJ in 2008 and now used by many guideline bodies, rates the quality of evidence separately from the strength of a recommendation. The core idea is that a reader should be told how much confidence the evidence deserves, not just what it found. Randomised trials in people start high. Observational data starts lower. Animal work does not tell you directly what happens in people at all. The library applies the same principle: the grade sits beside every summary so the summary cannot be read without it.
The how peptide research works page explains the steps from cell study to phase 3 trial, and why researchers study peptides covers what makes them interesting in the first place. The editorial policy sets out how pages are sourced and reviewed.
Regulatory status is part of the picture
Several compounds in the library are not approved for human use in the United States, and some are sold as research chemicals. The library says which is which. Policy changes too: the FDA's advisory committee review of several peptides is covered on the PCAC July 2026 page, and the research chemical and compounded glossary entries explain the terms. None of this is a judgement on any clinic, pharmacy or supplier. It is the regulator's status, stated plainly.
Who it is for
- Anyone who has been told about a compound and wants to know what the evidence actually is before talking to a clinician.
- People already on a protocol who want the contraindications and common combinations at the point where they log a dose.
- Clinicians and researchers who want a quick summary with the human and animal data kept apart.
What it does not do
The reported dosing ranges are what studies and the literature describe. They are not a recommendation, and the app never recommends a dose. Peptide AI does not sell, prescribe or recommend compounds.
How it connects to the rest of the app
Every compound in the library can go into the protocol planner and the dose log, with its reconstitution handled by the vial tracker. Remi cites the research behind each compound when it answers questions about your protocol. Each compound also has a room in the community. When you upload bloodwork, the markers are plotted against the weeks you were on each compound from the library.
Before you add a compound to a protocol, read its grade and contraindications in the library, and log it from there so the dose links back to the evidence. The app is on the download page.
Questions people ask
How many compounds are in the Peptide AI library?
57, including approved GLP-1 receptor agonists such as semaglutide and tirzepatide, investigational drugs such as retatrutide, and research peptides such as BPC-157 and TB-500.
What is the evidence grade?
A rating on every compound page of how strong the human evidence is. Where almost no human data exists, the page says so rather than filling the gap.
Are the dosing ranges recommendations?
No. They are ranges reported in the literature. Peptide AI does not recommend doses; that is for you and your prescriber.
Does the library say whether a compound is approved?
Yes. Regulatory status is part of each page, including compounds that are not approved for human use in the United States.
Can I see the library without the app?
Yes. The same compounds are on the website at peptideai.co/compounds.
Sources
- Every feature, in depth, Peptide AI
- Platform capabilities, Peptide AI
- How it works, Peptide AI
- Peptide AI on the App Store, Apple App Store
- GRADE: an emerging consensus on rating quality of evidence and strength of recommendations, BMJ, 2008
