Pointify

Voice-note app: record audio, get transcriptions (Whisper), and AI-generated bullet summaries (Ollama). Organize notes by pages and days with JWT auth and an optional Docker deployment.

Features

• Voice recording — Record in the browser; uploads stored and transcribed with Whisper
• Bullet digests — Ollama turns transcripts into bullet points (configurable model)
• Pages & days — Group recordings by page and date; reorder bullets via drag-and-drop
• Auth — JWT login, refresh tokens, optional admin panel and user management
• Production — Docker Compose (API + React frontend behind nginx), healthchecks, non-root user

Prerequisites

• Local dev: Python 3.13+, Node 22+, Ollama (for digests)
• Docker: Docker and Docker Compose (for production-style run)

Local development

Backend

cd backend
python3 -m venv .venv
source .venv/bin/activate   # Windows: .venv\Scripts\activate
pip install -r requirements.txt

Copy backend/.env.example to backend/.env and set at least:

• JWT_SECRET — random string for signing tokens
• OLLAMA_URL — e.g. http://localhost:11434
• OLLAMA_MODEL — e.g. your preferred model name
• DB_PATH — e.g. data/pointify.db
• CORS_ORIGINS — e.g. http://localhost:5173

Optional: ENABLE_DOCS=true for Swagger at /docs (local only).

Frontend

cd frontend
npm install

Run everything

From the project root:

./start.sh

• App: http://localhost:5173
• API: http://localhost:8000
• Docs: http://localhost:8000/docs (if ENABLE_DOCS=true)
• Default login: admin / admin — change on first login

Ensure Ollama is running (ollama serve) and the digest model is pulled (e.g. ollama pull <OLLAMA_MODEL>).

Production (Docker)

1. Copy .env.example (in project root) to .env and set values — especially DATA_PATH, WHISPER_CACHE_PATH, JWT_SECRET, OLLAMA_URL, and CORS_ORIGINS.
2. Run:

./run-prod.sh

Then point your reverse proxy (e.g. Nginx Proxy Manager) at the pointify-frontend container on port 80.

• Logs: docker-compose logs -f
• Stop: docker-compose down

Environment variables

Variable	Description
OLLAMA_URL	Ollama API base URL
OLLAMA_MODEL	Default model for bullet digests
JWT_SECRET	Secret for signing JWT tokens
DB_PATH	Path to SQLite database file
CORS_ORIGINS	Allowed origins (comma-separated)
LOG_LEVEL	Logging level (e.g. INFO)
MAX_UPLOAD_MB	Max upload size in MB
MAX_RECORDING_MINUTES	Max recording length in minutes
ENABLE_DOCS	Set true only for local dev (Swagger UI)

For Docker, also set DATA_PATH and WHISPER_CACHE_PATH for persistent storage.

Tests

• Backend: cd backend && pytest
• Frontend: cd frontend && npm test
• Lint / format: ruff and pre-commit in backend; ESLint/Prettier in frontend

Stack

• Backend: FastAPI, SQLAlchemy (SQLite), Whisper (faster-whisper), Ollama (HTTP), JWT + bcrypt
• Frontend: React 19, Vite, React Router, @dnd-kit (drag-and-drop)
• Deploy: Docker (Python image + nginx for static frontend), docker-compose