Diagora Assist

AI-powered clinical support system designed for real-world and low-resource settings, helping clinics and NGOs triage patients, detect risks, and deliver clear, explainable medical decisions faster

Comment

Inspiration

In many parts of the world, access to specialized medical care is limited. Frontline clinics, NGOs, and remote healthcare providers often operate under pressure, with limited resources, delayed diagnostics, and no access to multiple specialists.

Medical decisions in such environments are critical — and delays or uncertainty can cost lives.

We asked: What if AI could simulate a clinical team to support these environments?

This question led to the creation of Diagora Assist.

What it does

Diagora Assist is an AI-powered clinical support system designed for real-world and low-resource settings.

It simulates multiple medical specialists (Cardiology, Pulmonology, Neurology, Risk Analysis, and Lab Interpretation) working together to analyze patient cases.

Instead of providing a single answer, the system:

  • Analyzes symptoms and available patient data
  • Simulates structured clinical reasoning between specialists
  • Detects hidden risks and conflicting interpretations
  • Surfaces uncertainty instead of hiding it
  • Generates a clear, explainable, risk-aware clinical report

It helps clinics and NGOs triage patients faster and make safer decisions — even with limited data.

How we built it

Diagora Assist is built as a modular, scalable system inspired by real healthcare infrastructure:

  • A2A (Agent-to-Agent Communication): Enables structured interaction between AI specialists
  • FHIR-inspired Data Layer: Standardizes patient data for interoperability across systems
  • MCP Tools Layer: Adds modular capabilities like risk scoring, lab analysis, and imaging interpretation
  • RAG (Retrieval-Augmented Generation): Enhances reasoning using contextual medical knowledge
  • Custom Orchestrator: Coordinates agents, manages debate, and generates final insights

Frontend:

  • Next.js (App Router)
  • TypeScript
  • Tailwind CSS

Challenges we ran into

One of the biggest challenges was simulating realistic clinical reasoning instead of simple response generation.

We needed to:

  • Design meaningful interactions between multiple agents
  • Represent disagreement and uncertainty clearly
  • Keep the system understandable for non-expert users

Another challenge was balancing complexity with usability — especially for environments where time and clarity are critical.

Accomplishments that we're proud of

  • Successfully simulating multi-specialist clinical reasoning
  • Turning disagreement into a useful signal instead of a problem
  • Designing a system that works even with limited data
  • Creating a clean, intuitive UI for complex medical insights

What we learned

We learned that in healthcare, transparency and clarity are just as important as accuracy.

AI should not only give answers — it should explain reasoning, highlight uncertainty, and support better decisions.

We also learned that designing for real-world constraints (limited data, time pressure) changes how AI systems should be built.

What's next for Diagora Assist

  • Integration with real clinical systems and telemedicine platforms
  • Expanding specialist agents and diagnostic capabilities
  • Adding continuous patient monitoring and follow-up tracking
  • Supporting NGOs and field clinics in real-world deployments

Disclaimer

Diagora Assist is a clinical decision support prototype and does not replace licensed medical professionals.

Built With

  • a2a-communication
  • data
  • fhir-inspired
  • mcp-tools
  • multi-agent-ai-system
  • next.js
  • rag-(retrieval-augmented-generation)
  • tailwind-css
  • typescript
Share this project:

Updates