Skip to content

idea.md

Latest commit

 

History

History
51 lines (42 loc) · 2.42 KB

idea.md

File metadata and controls

51 lines (42 loc) · 2.42 KB
Raw

Prompt Prep | Precision Study Ecosystem

Problem

Students drown in unstructured PDFs and lecture notes. They lack the tools to quickly test their understanding or create structured review materials, leading to passive reading instead of active learning.

Solution

PromptPrep is a full-stack application that ingests educational documents (PDFs, text files), indexes them using RAG (Retrieval-Augmented Generation), and provides three AI-powered services:

  1. Context-Aware Q&A — Chat with your notes using an AI assistant that retrieves relevant passages.
  2. Automated Quiz Generation — Generate MCQ quizzes at varying difficulty levels from your materials.
  3. Flashcard Generation — Auto-generate term/definition flashcards for efficient review.

Key Features

  • Document Ingestion — Upload PDFs or text files; automatically parsed, chunked, and indexed.
  • Vector Search — Semantic similarity search powered by Supabase pgvector and Google Gemini embeddings.
  • Quiz Engine — AI-generated MCQs with explanations and scoring.
  • Flashcard API — Auto-generated term/definition flash cards.
  • RAG Chat — Ask questions and get answers grounded in your documents.
  • Interactive Demo — Try the full workflow without uploading a real document.
  • LLM Fallback System — Supports Gemini and Groq API keys with automatic failover.

Tech Stack

Layer Technology
Backend TypeScript, Node.js, Express.js
ORM Prisma (PostgreSQL)
Database & Vectors Supabase (PostgreSQL + pgvector)
AI/LLM Google Gemini API, Groq API (fallback)
Framework LangChain.js-compatible architecture
Frontend React, Vite, TypeScript
Styling Vanilla CSS (premium dark theme)

Design Patterns

Pattern Implementation
Strategy IDocumentParser → PDFParser, TextParser
Factory Method ParserFactory, GeneratorFactory
Template Method BaseContentGenerator (abstract class)
Singleton VectorStoreManager, EnvConfig
Repository BaseRepository<T> → concrete repos

Architecture

Controllers → Services → Repositories → Database (Supabase PostgreSQL)
                ↓                                ↓
        VectorStoreManager        →      pgvector (Similarity Search)
                ↓
          LLM (Gemini / Groq)