MediAI — AI-powered preliminary health assessment platform

MediAI is a modular web application that delivers fast, AI-driven preliminary health assessments from user-reported symptoms. The core prediction engine uses a supervised classifier (configurable; default: Support Vector Classifier) trained on symptom→disease mappings. Users submit comma-separated symptoms or dictate them via speech; inputs are normalized and encoded as a fixed-length binary feature vector before being scored by the preloaded model. Results include ranked diagnoses with probability/confidence scores and are augmented by five editable, CSV-driven content sections: Description, Precautions (four concise measures), Medications, Dietary Advice, and Workouts.

Architecture & responsibilities

• Model layer: model loading, prediction API, evaluation utilities (pluggable model support).

• Content/data layer: editable CSV lookups for condition metadata and guidance content.

• Input processing: text normalization, speech-to-text integration, vectorization into fixed-length binary arrays.

• Presentation layer: Flask + Bootstrap front end with a modern glassmorphism UI (translucent cards, gradient accents, smooth fade-ins).

• Deployment: production-ready under WSGI (Gunicorn) and containerizable with Docker.