How lakshay1509/hackhazard-25 works
Project Type & Purpose
• This is a full-stack, microservice-oriented application for real-time vocal health analysis.
• Core user touchpoints:
• Web client (React + Redux)
• Telegram bot (Python)
• Core processing:
• Node/Express API server
• AI microservice (Python model inference in Docker)
• Data store: MongoDB
• External services: Groq API, Cloudinary, Google APIs, hosting/CDN
Top-Level Component Breakdown
• Frontend “client” directory ➔ React (Vite), TailwindCSS, Redux
• Backend “server” directory ➔ Node.js + Express, controllers, routes, models, MongoDB driver
• AI microservice “ai” directory ➔ Python inference code, TensorFlow saved model, report generator, runs in its own Docker container
• Telegram service “telegram” directory ➔ Python bot, report generator, interacts with users via Telegram Bot API
Key External Dependencies & Hosting
• MongoDB Atlas or self-hosted MongoDB
• Cloudinary for file storage (audio file uploads, PDF reports)
• Groq API for low-latency voice/NLP inference
• Google VGGish embeddings
• Netlify (frontend hosting), Azure or Docker Swarm (server & AI containers)