A drug-centered multi-agent RAG system for knowledge retrieval, reasoning, and evidence synthesis across 70 curated biomedical resources.
Most biomedical QA systems stop at "retrieve a few passages and summarize them." Drug questions demand precision, provenance, and multi-hop reasoning.
Each skill ships with its own SKILL.md and example.py. The Code Agent reads both, learns native usage patterns, and generates query code dynamically.
70 curated drug resources organized into a navigable skill tree spanning DTI, ADR, DDI, PGx, repurposing, and more.
Turn free-form retrieval results into triples, subgraphs, ranked paths, and evidence-aware answers instead of simply stitching excerpts.
Built-in DuckDuckGo + PubMed search as a supplement for recent literature and external evidence beyond local resources.
Opinionated around drug-native tasks from resource organization to retrieval strategy, reasoning flow, and final answer structure.
Works with any OpenAI-compatible API: GPT-4o, Azure OpenAI, LLaMA via vLLM, Ollama, Together AI, and more.
Choose the right reasoning depth for your query — from fast retrieval to deep multi-hop graph reasoning.
Retrieve → Graph Build → Rerank → Respond → Reflect. Full evidence synthesis with optional iterative refinement and web search.
Retrieve and answer directly. Fast, efficient, and ideal for straightforward drug information queries.
Use only online search and literature retrieval. Perfect when you need the latest published evidence.
A multi-agent pipeline from query understanding to evidence-grounded answers.
User Query │ ▼ Retriever Agent ├── navigates the 15-subcategory skill tree ├── extracts key entities └── selects relevant resources │ ▼ Code Agent ├── reads SKILL.md + example.py ├── writes custom query code └── executes resource-specific retrieval │ ├──▶ SIMPLE ──▶ Responder ──▶ Final Answer │ └──▶ GRAPH ──▶ Graph Builder ──▶ Reranker ──▶ Responder ──▶ Reflector ──▶ optional Web Search ──▶ Final Answer
Organized across 15 drug-specific subcategories for comprehensive coverage.
| Category | # | Skills |
|---|---|---|
| DTI | 10 | ChEMBL, BindingDB, DGIdb, Open Targets, TTD, STITCH, TarKG, GDKD, Molecular Targets ×2 |
| ADR | 4 | FAERS, SIDER, nSIDES, ADReCS |
| Knowledgebase | 8 | UniD3, DrugBank, IUPHAR/BPS, DrugCentral, CPIC, PharmKG, WHO EML, FDA Orange Book |
| Mechanism | 1 | DRUGMECHDB |
| Labeling | 3 | openFDA Human Drug, DailyMed, MedlinePlus Drug Info |
| Ontology | 4 | RxNorm, ChEBI, ATC/DDD, NDF-RT |
| Repurposing | 6 | RepoDB, DRKG, OREGANO, Drug Repurposing Hub, DrugRepoBank, RepurposeDrugs |
| PGx | 1 | PharmGKB |
| DDI | 3 | MecDDI, DDInter, KEGG Drug |
| Toxicity | 4 | UniTox, LiverTox, DILIrank, DILI |
| Combination | 2 | DrugCombDB, DrugComb |
| Molecular Prop | 1 | GDSC |
| Drug-Disease | 1 | SemaTyP |
| Reviews | 2 | WebMD Drug Reviews, Drugs.com Reviews |
| Drug NLP | 7 | DDI Corpus 2013, DrugProt, ADE Corpus, CADEC, PsyTAR, TAC 2017 ADR, PHEE |
Up and running in minutes with just a working LLM config.
Raise resource density, retrieval fidelity, and evidence-grounded reasoning — all at the same time.