Overview
PreparationPalAI is a local-first study assistant for learning from your own documents. It combines retrieval-augmented generation with quizzes, flashcards, web scraping, and programming help so learners can turn notes and reference material into an interactive study workflow.
The application runs with a React frontend and FastAPI backend, using LangChain, LangGraph, ChromaDB, and Ollama. Documents stay on your machine by default, and the local model setup needs no paid API key.
Highlights
- Document Q&A with retrieved source context and citations
- Quiz preparation with generated questions, answers, and feedback
- Flashcards for active recall and spaced study sessions
- Local models through Ollama, with optional cloud-provider integrations
- Multiple orchestration modes using LangChain and LangGraph
- Monitoring stack with Prometheus, Loki, and Grafana configuration
Getting Started
git clone https://github.com/Markkimotho/PreparationPalAI.git
cd PreparationPalAI
pip install -r requirements.txt
ollama pull llama3.2:1b
ollama pull nomic-embed-text
streamlit run app/ui.py
The project also includes Docker Compose configuration for running the application and its supporting services together.
Technology Stack
| Component | Technology |
|---|---|
| Frontend | React, TypeScript, Vite |
| Backend | FastAPI, Python |
| AI orchestration | LangChain, LangGraph |
| Vector store | ChromaDB |
| Local models | Ollama |
| Observability | Prometheus, Loki, Grafana |