AgentHealth

• 

  Real-time healthcare agent network visualization with seamless data flow and API integration.

Inspiration

Healthcare data is often trapped in silos. While AI agents are revolutionizing healthcare, they struggle to communicate effectively with legacy systems and each other. We were inspired to build a unified platform where Model Context Protocol (MCP), Agent-to-Agent (A2A) communication, and FHIR standards converge to create a seamless healthcare data ecosystem.

What it does

AgentHealth is a comprehensive healthcare interoperability dashboard. It simulates a live hospital environment where:

Agent Network Topology: Users can visualize data packets flowing between AI agents (MCP, A2A, FHIR parsers) in real-time. Clinical Dashboard: Features a live ECG monitor with Normal, Tachycardia, and Bradycardia modes, alongside vital signs monitoring. AI Health Assistant: An intelligent chatbot for patient queries, diagnosis suggestions, and FHIR resource explanations. FHIR Builder: An interactive tool to create, validate, and download FHIR R4 compliant resources (Patient, Observation, etc.).

How we built it

We built this using Vanilla JavaScript for high performance and Tailwind CSS for a responsive, modern UI. The frontend simulates backend agent behaviors, including data packet routing and clinical decision support logic. We focused heavily on SVG animations for the ECG monitor and network visualization.

Challenges we ran into

Simulating real-time ECG waveforms with different heart rate conditions (Tachycardia/Bradycardia) using pure SVG paths was complex. Ensuring the FHIR resources generated were R4 compliant also required strict validation logic.

Accomplishments that we're proud of

We are proud of the "Live ECG Monitor" which mimics real ICU equipment visuals and the "Achievement System" that adds gamification to the user experience. The platform successfully demonstrates how agents can collaborate to process patient data.

What we learned

We deepened our understanding of healthcare interoperability standards (FHIR R4) and the emerging protocols like MCP and A2A. We also learned advanced CSS animations for creating immersive UIs.

What's next for AgentHealth

The next step is to integrate real Large Language Models (LLMs) for the AI Assistant and connect to actual FHIR sandboxes for live data testing. We aim to add more clinical modules like radiology integration.

Built With

• css
• css3
• fhirr4
• html5
• javascript
• svg
• tailwind