Minimal RAG-powered PDF learning and practice platform.
Upload any document to stream answers, inspect extracted topic structures, and practice with auto-generated multiple-choice questions.
- Q&A: Streamed LLM responses backed by document context.
- Topics: Automatic detection of headings and sections.
- Practice: Topic-targeted MCQ quizzes with explanations.
- Search: BM25 keyword matching + vector similarity.
- Backend: FastAPI, PyMuPDF, Groq, Google GenAI, Pinecone
- Frontend: React (Vite)