Clinical trials
Studies on ClinicalTrials.gov by drug, condition, sponsor, phase or status, with their design, eligibility, sites and results.
Research clinical trials and drugs in your native language
ClinLit AI answers questions about clinical trials, protocols, drugs and the published literature from public data, and cites every claim back to its source record: an NCT ID, a PMID, an FDA application or label, or a section of a guideline.
FDA drug data
Follow a drug from its trials to its approval, patents and exclusivity, its label, recalls and shortages, and the adverse event reports FDA receives.
Published literature
PubMed abstracts, linked to the trials they report on.
Protocols and guidance
Protocols, analysis plans and consent forms, and the ICH E6(R3) GCP guideline.
Charts and maps
Answers as charts, heatmaps and maps of trial sites, with the underlying data one click away.
From the blog
All posts →How to design a phase 2b MASH trial: lessons from 219 trials and their protocols
Placebo response from 2% to 33%, screen failure up to 87%: what 219 phase 2 MASH trials and six protocols show about endpoints, sample size and biopsies.
How we researched it →Eli Lilly R&D profile, November 2026: pipeline, trials and peer comparison
Eli Lilly's clinical pipeline as of 7 November 2026: 240 active trials, its late-stage drugs, readouts due, FDA approvals, and how it compares with four peers.
How we researched it →Cleaning openFDA data: how we linked FAERS, drug labels and approvals to trials and PubMed
What it took to turn openFDA's FAERS reports, drug labels and approvals into linked, searchable data: the problems we hit, the fixes, and how we checked them.
- clinical trials
- 430K+
- PubMed articles
- 17M+
- FDA drug labels
- 90K+
- serious adverse event reports
- 7.8M
From the creator
I’m a solo developer with a background in AI, consulting and pharma R&D. I built ClinLit AI for people who know clinical research inside out but don’t have AI and data engineering teams or an enterprise budget.
It’s in a pilot phase, running on modest computing power, so some questions may take a while. Give it a try, and please tell me what it gets wrong.
Your questions are sent to a third-party AI model, so please don't enter personal or other sensitive information. Like any AI, it can make mistakes, and its answers are not medical, investment or regulatory advice. What's logged and how to check answers →