Work
The current, synthesized view of why a trial stopped, whether two names are the same molecule, and what remains unknown. Built for BD, clin-dev search & eval, tech-transfer, and NIH-adjacent review — so the next development decision is faster and cited.
Corpus empty — ingest not running. Live opportunities appear after `npm run ingest` or an authenticated cron tick. Paste an NCT to open a program from ClinicalTrials.gov now.
Start here
Click a live opportunity, or paste an NCT. That is the only first move. Identity, why it stopped, science still open, and the memo update from that record.
ResurrectAI
Public analysis. Grounds on ClinicalTrials.gov and PubMed, then ranks addressable-failure hypotheses. Empty candidates if the case is weak. Not medical advice.
A trial can stop because the drug failed, or because enrollment, money or a company event intervened. Only the first answers the scientific question. The classifier reads the sponsor’s own filing; if they filed nothing, we mark it unknown.
MK-3475, pembrolizumab and Keytruda are one program. Pembrolizumab and nivolumab are not — they share a -mab suffix, not a target. Click an alias on a loaded asset to see the merge.
A packet for IC or a partner: why it stopped, what science is still open, who ran it, program size. Every line is cited or marked unknown. It does not claim the drug works.