Overview
MediVision AI is a clinical assistant platform designed to demonstrate how AI can assist in medical triage and preliminary diagnostic evaluation. It allows users or healthcare practitioners to upload clinical images along with voice-recorded symptoms or text notes, which are processed by vision-language models to output a detailed medical report.
Core Features & Functionality
Multimodal Diagnostic Inputs
Medical Image Analysis: Supports drag-and-drop file uploads or direct webcam capture for evaluating skin lesions and visual clinical conditions.
Voice & Symptom Recording: Built-in audio recorder with canvas waveform visualization to capture spoken patient symptoms, alongside standard text inputs.
Sample Cases: Pre-loaded clinical sample scenarios for quick testing and demonstration.
AI Diagnostic & Voice Engine
Transcription & Assessment: Transcribes patient audio, evaluates image features, and outputs a structured clinical assessment detailing differential diagnoses and treatment guidelines.
Doctor Voice Response: Generates audio playback of the assessment using high-fidelity text-to-speech engines (ElevenLabs Turbo v2 with Web Speech fallbacks).
Backend & Integration Options
Database History: Connects with MongoDB Atlas (Cluster0) to store and retrieve past diagnostic sessions and history.
Flexible API Configuration: Supports custom API keys (such as Groq for inference and ElevenLabs for voice) or standalone demo mode running directly in the browser.
