Neutral, cited, evidence-graded

See what the science actually says about GLP-1 drugs.

A neutral reference library of what has been observed across GLP-1 and incretin drugs, marketed and investigational. Every finding carries its source and a grade for how strong the evidence is.

Semaglutide nausea Tirzepatide heart rate Retatrutide Muscle loss Appetite Liver / MASH

Browse the library Compare drugs Create a free account How it works

Browse is public, no login. Account and MCP access are scaffolded for early testers.

Tirzepatide MARKETED

blood-pressure Decrease

High evidence

Tirzepatide reduced 24-hour ambulatory systolic BP versus placebo (SURMOUNT-1 ABPM substudy); effect partly weight-mediated.

de Lemos JA et al. (SURMOUNT-1 ABPM), Hypertension, 2024;81(4):e41-e43 |human-primarydouble-blind RCT Source

PopulationN=600 adults BMI >=27, BP <140/90 on stable antihypertensives, SURMOUNT-1 substudy; 36 weeks; RCT.

Fundingindustry - Eli Lilly

Comparatorsplacebo

651 results | 18 ongoing records | 471 sources | every one cited and graded | not medical advice

GLP-1 information is loud. Evidence is quiet.

Press releases read like pivotal trials. A rodent study gets the same headline as a large randomized trial. Forums, sellers, and hype all pull in different directions, and the newest agents have the least settled science of all.

So this library does one thing, carefully: it shows each finding with its source and how strong that source is, and it never pretends to more certainty than exists.

Every finding is cited.

Source, channel, and date ride with each one. Strip the provenance and it is not here.

Every finding is graded.

A four-level robustness grade separates pivotal evidence from press releases, abstracts, and early signals.

No verdicts. By design.

Disagreements are left standing side by side. It records what has been observed. It does not tell you what to do.

The anatomy

What one finding looks like.

Every entry is built the same honest way. This one teaches you to read the whole library.

Semaglutide MARKETED

Specific groups of people Increase

Moderate evidence

In a PCOS RCT, adding semaglutide to metformin raised the natural pregnancy rate (35% vs 15%) in anovulatory women.

Semaglutide+metformin in PCOS (open-label RCT, n=100), Reprod Biol Endocrinol 2025 |human-primaryopen-label RCT Source

PopulationPCOS open-label RCT (NCT not stated; Reprod Biol Endocrinol 2025): 100 overweight/obese women with PCOS (Rotterdam criteria), prospective randomised controlled open-label trial of semaglutide 1 mg weekly plus metformin vs metformin alone, 16 weeks.

Fundingacademic/government - Chinese government and institutional grants (Chongqing Health Commission/Science & Technology Bureau; NSFC; not industry)

Drug and status

Marketed and investigational records are visibly separated. Investigational means unapproved or pipeline context, never endorsed.

Neutral effect domain

Domains organize evidence using neutral labels such as blood pressure, appetite, body composition, and pharmacology.

Evidence grade

The grade rates source strength, not whether a drug is good or whether an effect improves health.

Provenance

Citation, source type, maturity, funding, and reader-facing limitations stay attached to the finding.

Expandable details

Tester-facing cards expose funding, population, limitations, and caveats without leaking internal audit notes.

Side by side

Lay two drugs side by side.

Pick any two agents and read the differences directly, domain by domain, each finding graded and cited. Where the evidence is thin on one side, you see that too.

Compare drugs

Tirzepatide

MARKETED

Semaglutide

MARKETED

safety-other

Tirzepatide High evidence Tirzepatide is contraindicated in the US for personal/family history of MTC or MEN 2 (also a boxed warning) and serious... Mounjaro (tirzepatide) US Prescribing Information, Eli Lilly; FDA/MHRA/TGA advisories

Semaglutide High evidence NET VERDICT (early-worsening diabetic retinopathy): SUSTAIN-6 found retinopathy complications significantly MORE frequent with... Marso SP et al. (SUSTAIN-6), NEJM, 2016

special-populations

Tirzepatide High evidence The EU SmPC states tirzepatide could be considered for use during breast-feeding. Mounjaro (tirzepatide) EU Summary of Product Characteristics (EMA)

Semaglutide High evidence The EU SmPC states semaglutide should not be used during breast-feeding. Ozempic (semaglutide) EU Summary of Product Characteristics (EMA)

comparative-efficacy

Tirzepatide High evidence DIRECT head-to-head CVOT (SURPASS-CVOT, T2D + ASCVD). Tirzepatide (up to 15 mg) was NON-INFERIOR to dulaglutide 1.5 mg for... Nicholls SJ et al. Cardiovascular Outcomes with Tirzepatide versus Dulaglutide in Type 2 ...

Semaglutide Moderate evidence DIRECT head-to-head (SUSTAIN 7, T2D). Semaglutide was superior to dulaglutide at matched low and high doses for both HbA1c and... Pratley RE, Aroda VR, Lingvay I, et al. Lancet Diabetes Endocrinol 2018 (SUSTAIN 7)

Coverage honesty

Honest about what is thin.

The evidence is uneven, and the library shows it plainly. The marketed leaders carry most of the high-grade trials. Many newer agents rest on early-phase, conference, or press evidence.

The same honesty applies to hype. A topline press release is recorded, and graded at the floor.

Retatrutide INVESTIGATIONAL

Heart rate Increase

Very low evidence

In isolated mouse right atria, retatrutide itself exerted positive chronotropy that was antagonised by the GCGR antagonist adomeglivant, potentiated by rolipram (PDE4 inhibition) and abolished by the PKA inhibitor H89 - but NOT weakened by propranolol; the...

Neumann J et al., Naunyn Schmiedebergs Arch Pharmacol 2025;398(12):17147-17160 |rodent-primarypreclinical (rodent) Source

Full findingIn isolated mouse right atria, retatrutide itself exerted positive chronotropy that was antagonised by the GCGR antagonist adomeglivant, potentiated by rolipram (PDE4 inhibition) and abolished by the PKA inhibitor H89 - but NOT weakened by propranolol; the authors conclude retatrutide excites beating rate via GCGR, signalling through cAMP/PKA.

Populationisolated mouse right atria (and left atria for force of contraction)

Fundingacademic/independent analysis (manufacturer not study sponsor; per-study COI not individually audited)

Comparatorsadomeglivant; rolipram; H89; propranolol

For the technical

Connect it to your own AI.

The whole library is available as a connector for AI clients that support MCP. Point your own assistant at it and every answer can be grounded in cited, graded findings rather than the open web. The connector serves data only. It runs no model and invents nothing.

The connector is in early access. Testers get it free in exchange for feedback.

Apply to be a tester

If you are researching these drugs

See the evidence behind the headlines, at a grade you can weigh.

If you are a clinician or researcher

A fast, cited, neutral cross-reference, every claim traceable to source.

If you build with AI

A provenance-bearing data source for your own tools.

Early access

Help shape the connector. Test it free.

The MCP connector is in early access. We are looking for a small group of testers to use it with their own AI setup and tell us what works and what does not.

OKUse it with your own MCP-capable client

OKSend usability feedback now and then

OKFree while in early access; premium features may be chargeable in future

Apply to be a tester

Questions, answered plainly.

No. It is a research reference. It does not diagnose, recommend, dose, or tell anyone what to take.

It means the source behind that finding can bear more weight. It does not mean the drug is good, appropriate, safe for a person, or recommended.

Because their evidence is often the hardest to judge. They are labelled as investigational, graded, and never endorsed.

Yes, through the MCP connector during early access. The connector serves cited corpus data; it does not generate medical advice.

Read the evidence for yourself.

Browse the library Apply to be a tester