[UNTESTED] [Markkimotho/PreparationPalAI] 1 endpoint(s)
Gemini
API Proxy/Gemini
2,158 characters
Status: [UNTESTED]
Endpoints found:
http://localhost:3100/loki/api/v1/label/compose_service/values
Source: https://github.com/Markkimotho/PreparationPalAI
# PreparationPalAI — RAG · Quiz · Web Scraping · Tool Calling · Observability
A fully local AI study tool that turns any document or web topic into interactive quizzes. Upload your PDFs, paste a URL, or just name a topic and it scrapes the web — then generates MCQ, True/False, and Short Answer questions, grades your answers with feedback, and explains concepts on demand. Everything runs on your machine via Ollama. No API keys. No cloud costs.
The primary interface is a **React** single-page app with a collapsible navbar — **Quiz Prep**, **Study Agent**, **Programming Assistant**, **Regular Chat**, **Flashcards**, and an in-app **Guide** — served by a **FastAPI** backend that streams responses token-by-token. Upload PDFs **or images** (OCR'd automatically); your documents are reachable from every chat section. The legacy Streamlit UI (Multi-Model Analysis + Insights) remains available.
Built on React + FastAPI + LangChain + LangGraph + ChromaDB + Ollama.
---
## Quick Start
### Docker (recommended)
No local Python, Node, Ollama, or model setup required. One command starts everything.
```bash
git clone <repo-url> && cd PreparationPalAI
docker compose up --build
```
On first boot, Docker pulls `qwen2.5:7b` (~4.7 GB) and `nomic-embed-text` (~274 MB) automatically, builds the React SPA, and starts the API. Open [http://localhost:8000](http://localhost:8000) once the models are ready. Subsequent starts are instant — models are cached in a Docker volume.
### Local development
Two processes: the FastAPI backend and the Vite dev server (which proxies `/api` to the backend).
```bash
# 0. Install Ollama and pull the required models
brew install ollama && brew services start ollama # macOS
# curl -fsSL https://ollama.com/install.sh | sh # Linux
ollama pull qwen2.5:7b # chat, quiz, code, agent
ollama pull nomic-embed-text # required for all RAG operations
# 1. Backend
git clone <repo-url> && cd PreparationPalAI
python -m venv .venv && sourc