Agents-MCP-Hackathon/medication-interaction-agent

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

💊 Medication Interaction Checker

An intelligent assistant that helps users check for potential drug-drug interactions from free-form natural language descriptions. Built using smolagents and powered exclusively by Anthropic’s Claude Opus 4, this app demonstrates how Claude’s high reasoning ability can support clinical safety tasks in real-time.

🧠 How It Works

Users enter messages like:

• —"I'm taking warfarin and metformin. Is there any interaction?"

Claude Opus 4 handles everything:

• —medication names from natural text

• —Identifies active ingredients

• —Queries the RxNav API to get known interactions

• —Summarizes the clinical relevance in clear, human-readable language

🔍 Multi-Agent Architecture

• —Manager Agent: Plans, delegates, and summarizes

• —Sub-Agent: Handles ingredient extraction and interaction checking

All tools and agents are built on Claude Opus 4, showing its versatility across planning, extraction, API interpretation, and summarization

🛠️ Tech Stack

• —🤖 smolagents framework

• —🧠 anthropic/claude-opus-4-20250514 (via LiteLLM)

• —🌐 RxNav drug interaction API

• —🖼️ Gradio UI with a conversational agent interface (gr.ChatInterface)

🚀 Why Claude Opus 4?

I was captivated by this statement from Anthropic:

Claude Opus 4 is our most intelligent model to date, pushing the frontier in coding, agentic search, and creative writing.

In high-stakes domains like healthcare, Claude's reliability and interpretability make it the ideal choice.

🎥 Video Overview

Watch the video overview here — explains how the app works, key features, and real-world use cases.

🏁 Try It

Enter any natural sentence describing your medications and check for potential interactions instantly.

👨‍⚕️ Example Prompts

"I take amlodipine in the morning and ibuprofen sometimes. Should I worry?"

"Prescribed metformin and glipizide. Safe combo?"