🏥 HealthDecoded

Decoding complex health events into clear, actionable advice for everyone—kids, teachers, parents, and doctors.

🌟 Inspiration

The idea for HealthDecoded was born from a personal experience working with diabetes clinics in France. I witnessed firsthand how children with diabetes and their teachers often panicked during hypoglycemic episodes due to unclear explanations. Parents received stressful alerts without always understanding the root cause, and doctors lacked context to adjust treatments effectively.

Goal: Create a tool that decodes glucose events in real-time, adapting explanations to each stakeholder (child, teacher, parent, doctor) using:

• Thryve API for context (meals, sports, glucose levels).
• LLM (Mistral) to generate personalized messages.
• Simple interface (Streamlit) for a powerful demo.

🛠️ How I Built the Project

1. Technical Architecture

graph TD
    A[Thryve API] --> Context Data| B[Python Backend]
    B -->|Prompt + Context| C[Mistral LLM]
    C -->|Tailored Messages| D[Streamlit UI]
    D --> E[User: Child/Teacher/Parent/Doctor]

Built With

• elevenlabs
• mistral
• python
• thryves