How akash-selvaraj/microbert works
Project type and purpose
• This is a full-stack web application called “MicroBert” composed of:
• A Python-based backend (FastAPI served by Uvicorn)
• A React/Next.js frontend (TypeScript, TailwindCSS)
• Purpose: Provide a chat interface (presumably powered by a BERT-style model) in the browser, with the backend handling AI inference and API requests.
File structure
• Top-level:
• microbert_backend/ – Python service with app.py, requirements.txt, README
• microbert_frontend/ – Next.js app with pages under src/app, configuration files
• Frontend patterns:
• Next.js file-based routing in src/app/chat/page.tsx and src/app/about/page.tsx
• Reusable UI components in src/app/components
• CSS via Tailwind (tailwind.config.ts, globals.css)
• Backend:
• Single entry point app.py exporting a FastAPI app instance
• Dependency list in requirements.txt
• No database directories—likely stateless or in-memory AI calls; any state is ephemeral or managed externally.
README for additional insights
• Backend: run uvicorn app:app --host=0.0.0.0 --port=8000
• Frontend: run Next.js via npm start (defaults to port 3000)