MoneyPrinterTurbo
Turns a topic into a finished short video — script via LLM, stock or AI-generated footage, TTS, subtitles, FFmpeg assembly. 120k+ stars, MIT.
by Harry
MoneyPrinterTurbo takes a topic or keyword and runs the whole pipeline — script from an LLM (OpenAI, Claude, Gemini, DeepSeek, Qwen, Ollama and more), footage from Pexels, Pixabay or Coverr, or AI video generation (MiniMax, Seedance), voice via Edge TTS or ElevenLabs, subtitles from TTS timestamps or local faster-whisper, then FFmpeg assembly in 9:16, 16:9 or 1:1. WebUI (Streamlit), REST API (FastAPI) or CLI, same core. Docker with CPU/GPU variants; uv sync --frozen on macOS/Linux, one-click .exe on Windows. Config lives in config.toml.
I keep it as reference for wiring an LLM-to-video pipeline end to end without hand-building the footage sourcing or subtitle sync myself. MIT. Source harry0703/MoneyPrinterTurbo.