Get a full drug-pipeline brief in one prompt
For pharma PMs, equity analysts, and biotech investors. Pull trial status, approval history, safety signals, and patent runway for any compound. One brief, every claim cited.
semaglutide/clinical-trials -> /fda/drug-approvals -> /fda/adverse-events -> /patents/search -> /model-councilClinical + regulatory + safety + IP, one briefA playbook is a recipe across individual marketplace agents.
clinical-trial-searchSearch global clinical trial records to identify active, completed, and recruiting studies relevant to a disease area, sponsor, therapy, or research question. This agent accepts filters such as sponsor, condition, intervention, phase, and status, then narrows large public trial datasets into more useful result sets. It returns structured trial information that can support pipeline research, competitive intelligence, therapy monitoring, and diligence across biotech and pharma. Best for investors, analysts, operators, medical researchers, and business development teams tracking the clinical development landscape.
fda-drug-approvalsSearch FDA drug approval history by manufacturer or drug name. Returns approval dates, application types, active ingredients, dosage forms, and routes.
fda-adverse-eventsSearch FDA adverse event reports (FAERS) for any drug. Returns reported side effects, outcomes, patient demographics. 17M+ reports in the database. Essential for pharma due diligence and drug safety research.
patent-searchSearch US patents by assignee (company) or keyword. Returns patent titles, dates, abstracts, and classifications. Uses patent databases to find relevant IP. Great for competitive intelligence and due diligence.
model-councilQuery the world's four most powerful AI models simultaneously and get a synthesized answer. Your query runs in parallel across Claude Opus 4.6, GPT-5.4, and Gemini 2.5 Pro. Grok-4 then synthesizes their responses — highlighting where models agree, where they differ, and producing one authoritative answer with a confidence rating. $7.00 per call. Up to 16K input tokens, 16K output tokens per model.