MorphOS

Inspiration

In a medical crisis, software shouldn't be a hurdle; it should be an immune response. We were inspired by the rapid evolution of outbreaks like the COVID-19 surge in Yaoundé, where traditional, rigid ERP systems simply couldn't scale or adapt to changing protocols. We wanted to build a system that moves at the speed of thought, not the speed of a development cycle.

What it does

MorphOS is a Generative ERP that transforms hospital management into an intelligent, evolving entity.

• Generative Dashboards: Admins describe a crisis need in natural language, and the system instantly codes a functional React interface with live geospatial analytics.
• Napkin-to-UI: Doctors can sketch a new triage form on paper; MorphOS uses Gemini 3 Pro Vision to digitize it into a working clinical database entry in seconds.
• Hands-Free Scribe: A voice-activated clinical capture system designed for doctors in PPE, allowing for real-time data entry and verbal corrections.
• Agentic Operations: The system understands its own API, allowing it to autonomously simulate surges and suggest resource re-allocation via Tool Calling.

How we built it

• The Brain: Gemini 3 Pro handles the reasoning, vision, and multimodal inputs (voice/sketches).
• The Skeleton: A modular React frontend that supports dynamic component injection for the Generative UI.
• The Data Layer: We integrated the Google Sheets API as a lightweight, highly accessible database for rapid resource tracking and financial auditing.
• Agentic Logic: We implemented Tool Calling patterns that allow the AI to query internal databases and trigger external workflows (like PDF generation and email routing).

Challenges we ran into

• Deterministic UI Generation: Ensuring that a "Generative UI" isn't just pretty, but functional. We had to create strict prompt-to-code schemas to ensure the AI-generated React components didn't break the application state.
• Multimodal Latency: Optimizing the transition from a hand-drawn "Napkin" sketch to a structured JSON schema quickly enough for a high-stress medical environment.
• PPE Audio Clarity: Filtering out the muffled audio of a doctor speaking through a respirator or Hazmat suit to ensure the Hands-Free Scribe remained accurate.

Accomplishments that we're proud of

• Zero-Code Adaptation: We successfully demonstrated a shift from "Legacy Code" to "Generative Code," where a user with no technical background could build a crisis dashboard in under 30 seconds.
• The Vision Pipeline: Seeing a crude hand-drawn triage form become a live, validated digital form with active data fields felt like magic.
• Contextual Awareness: The system doesn't just store data; it "understands" the hospital layout and resource constraints.

What we learned

• The Power of Agency: We learned that an ERP shouldn't just be a database; it should be an Agent capable of reasoning through its own architecture.
• Empathetic Engineering: Designing for doctors in Hazmat suits taught us that the most sophisticated UI is sometimes the one you don't have to touch at all.

What's next for MorphOS

• Multi-Lingual Edge Cases: Fine-tuning the voice capture for the specific dialectic nuances found in Yaoundé and surrounding regions.
• Predictive Logistics: Using historical surge data to allow the AI to "pre-generate" UI modules before an admin even asks for them.
• Offline First: Implementing localized LLM instances to ensure MorphOS remains functional even during internet outages in remote regions.

Built With

• react
• typescript