mehedihasankanon/EchoDerm-AI ? reverse-engineered prompt

Build me a web app called EchoDerm AI that helps a rural clinic quickly tell the difference between measles and dengue from two things at once, one rash photo and a short cough recording.

I want a simple patient friendly upload screen in the frontend, with clear buttons to add an image and audio clip, then a results page that shows the likely diagnosis, a confidence score, what the AI noticed in the rash and cough, and a few next steps a clinician can follow. The backend should accept both files together, send them through the AI analysis, and return structured results in JSON. Please also save each diagnosis record so it can be reviewed later, and include a basic health check endpoint.

Use the existing React and TypeScript frontend and a Python FastAPI backend, and wire it up so it runs locally with a clean modern UI. If you need current setup details for any APIs or libraries, look them up online.