NEWT
daily tech news summaries in a reel-style format
Project metadata
Overview
A web application delivering daily tech news summaries in an engaging full-screen reel-style format, similar to YouTube Shorts or Instagram Reels. Uses AI-powered summarization so users can catch up on tech developments in a few minutes instead of browsing multiple sources.
Retrieval pipeline
NEWT runs a full end-to-end RAG system. Articles are embedded and stored in ChromaDB; retrieval over that vector store grounds both the daily summaries and the AI chat assistant. MongoDB backs user data, reading statistics, streaks, and the social graph.
Architecture
The mock server provides test data without requiring external APIs during development.
API
Requires Node.js v18+ and Python 3.12+. Frontend on :5173, backend API on :8000. A GEMINI_API_KEY is required in backend/.env.
Full stack
React · Vite · FastAPI · Python · Gemini API · ChromaDB · MongoDB · JWT
Results
FastAPI + React + Gemini + JWT | Yes | |
Mock dev server for offline development | Yes | |
ChromaDB vector search | Yes, full end-to-end RAG system | |
MongoDB | Yes |