How venkat2305/learnscribe works
Project type and purpose
• This is a full‑stack, single‑page web application (“Learn‑Scribe”) built with React+TypeScript on the frontend and FastAPI (Python) on the backend.
• Purpose: Let users generate and attempt AI‑powered quizzes and summaries from YouTube videos, articles, PDFs, etc., persisting results in MongoDB, and leveraging external LLM services (Gemini, Groq, etc.).
File structure
• Top‑level “frontend” folder: Vite + React app, pages (Auth, Dashboard, Quiz, Summary), Redux Toolkit store, UI components, service clients.
• Top‑level “backend” folder: FastAPI app organized into config, routers (auth, quiz, summary), Pydantic models, service layer (AI integrations, article extraction, YouTube transcript), utils (authentication helpers, quiz & summary generators), and async MongoDB connector.
• Deployment: Dockerfiles in frontend & backend, docker‑compose.yml to orchestrate services. CI via .github/workflows.
README for additional insights
• Uses Pydantic BaseSettings for config, Motor AsyncIOMotorClient for non‑blocking MongoDB access, OAuth2PasswordBearer for JWT auth, CORS middleware configured for secure cross‑origin calls.
• External AI services abstracted in a “llm_factory” and per‑provider modules (gemini.py, groq.py).
Elements
• Clear labels: “React SPA”, “FastAPI Routers”, “LLM Factory”, “MongoDB”
• User → SPA → API
• API → Services → external AI → Services → DB → API → SPA → User
• Blue rectangle: Frontend
• Green rectangle: Backend components (API, Services, Utils)
• Purple clouds: External AI & data sources