Notecast

Turn documents into a two-host podcast and chat with them. React frontend, FastAPI backend, Ollama for language, Edge TTS for audio.

Run with Docker

You need Docker and Docker Compose.

docker compose up --build

Then open http://localhost:3000.

The first start downloads the Ollama image and pulls llama3.2. That can take several minutes. Later starts reuse the volume.

API docs stay at http://localhost:8000/docs.

Stop with Ctrl+C, or docker compose down. Uploads, podcasts, and the database live in the backend_data volume. The model lives in the ollama volume.

Optional .env

cp .env.example .env

Variable	Default
SECRET_KEY	change-me-in-production
OLLAMA_MODEL	llama3.2
TTS_ENGINE	edge
TTS_VOICE_FEMALE	en-US-AriaNeural
TTS_VOICE_MALE	en-US-GuyNeural

Use it

1. Register / log in
2. Create a project
3. Upload a PDF or DOCX
4. Chat with the documents, or generate a podcast in Studio

Repo layout

backend/     FastAPI, TTS, Ollama client
frontend/    Vite + React app
docker-compose.yml

Run without Docker

# Ollama
ollama pull llama3.2

# API
cd backend
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
cp .env.example .env
python app.py

# UI
cd frontend
npm install
cp .env.example .env
npm run dev

UI: http://localhost:5173
API: http://localhost:8000

License

MIT