QSong-github

QSong-github / DrugClaw

Public

🦀 Agentic RAG for drug intelligence · 57 skills · 15 task categories · DTI · ADR · DDI · PGx · Repurposing · Powered by LangGraph

qsong-github.github.ioDrugClaw drug-discovery knowledge-graph llm multi-agent rag
59
2
100% credibility
Found Mar 18, 2026 at 60 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

DrugClaw is an AI-powered tool that gathers and analyzes drug information from dozens of specialized medical databases to answer questions about targets, reactions, interactions, and repurposing.

How It Works

1
🔍 Discover DrugClaw

You hear about DrugClaw, a smart helper for checking drug details like side effects or targets, perfect for safe choices.

2
📦 Get it ready

Download the folder and add a few easy helpers so it works on your computer.

3
🔑 Link your AI friend

Share a simple note with your preferred AI service, like a password, so DrugClaw can think deeply.

4
🚀 Try the quick demo

Run the demo to see it answer a sample question about a common drug, pulling facts from trusted sources instantly.

5
💭 Ask your question

Type your real question, like 'What side effects does this medicine have?', and pick simple or detailed thinking.

6
📊 Review the answer

Get a clear report with evidence from medical databases, targets, risks, and sources you can trust.

✅ Make informed choices

Now you understand your medicine better, feeling confident about safety, interactions, or new uses.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 60 to 59 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is DrugClaw?

DrugClaw is an agentic RAG system built in Python with LangGraph for querying drug intelligence across 70 curated resources in 15 categories like DTI, ADR, DDI, PGx, and repurposing. It handles complex queries generic assistants fumble—drug targets, adverse reactions, interactions, mechanisms—via a CLI like `drugclaw run --query "imatinib targets?"` in modes like GRAPH for multi-hop reasoning, SIMPLE for quick hits, or WEB_ONLY for literature. Users get traceable, evidence-synthesized answers from native resource queries, not flattened summaries.

Why is it gaining traction?

This agentic RAG architecture stands out from basic RAG by using code agents to query each resource natively, building knowledge graphs for reasoning, and reflecting on evidence gaps—far beyond agentic RAG vs RAG baselines in open source GitHub repos. Devs dig the skill tree navigation, web fallback for fresh data, and LangGraph workflows that feel like agentic GitHub Copilot for biomed. With 57 implemented skills, it's a ready agentic RAG pipeline for drug tasks, hooking those tired of manual PubMed dives.

Who should use this?

Biomedical engineers querying drug-target interactions or ADRs for discovery pipelines. Pharmacologists checking DDI risks or PGx guidelines before trials. ML devs building agentic workflows need precise, multi-source drug intel without scraping.

Verdict

Worth starring for agentic RAG LangGraph fans in biomed—CLI demos work out-of-box with OpenAI keys—but at 59 stars and 1.0% credibility, it's early; grab local resources from HF mirrors first. Solid niche prototype, not production-ready yet.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.