Drug Discovery
LiveDrug Discovery Agent automates end-to-end drug-target prioritisation with a LangGraph orchestrator, MCP-connected biomedical databases, deterministic scoring with conflict detection, and optional human-in-the-loop plan and review gates.
Drug Discovery highlights & features:
- LangGraph pipeline spanning planning, parallel collection, normalisation, scoring, evidence graphs, and synthesis
- MCP data layer querying DepMap, Pharos, Open Targets, and Europe PMC in parallel
- Deterministic multi-source druggability scoring with conflict severity tiers
- Five-layer memory system for reproducible, auditable research runs
- Optional plan approval and review gates before dossier acceptance
- FastAPI + Next.js research workbench with live SSE progress and follow-up Q&A
A production-style biomedical research agent that turns multi-database target assessment into a reproducible, traceable dossier in minutes.
Stack used
LangGraphFastAPIPythonNext.jsReactTypeScriptDocker