MoneyPrinterTurbo: one topic in, a finished short video out

What it is (and what's really free)

• The repo: github.com/harry0703/MoneyPrinterTurbo, about 141k stars, MIT license.
• Script (AI model): free if you run a local model with Ollama. Cloud models (OpenAI, Moonshot, etc.) need a paid API key.
• Voice: Edge TTS is free and needs no API key (official README).
• Clips: your own videos (free), or stock footage from Pexels / Pixabay with a free API key.

The "under 2 minutes per video" in the reel is the creator's own test on his machine. Your speed depends on your computer and on the AI model you pick. Star counts change: check the repo today.

What you need first

• Python 3.11 or later (the README recommends 3.11).
• uv, the Python package manager the project uses: install guide.
• Git, to download the project.
• Ollama (for the free local AI model): ollama.com/download.

Download and install

In a terminal:

git clone https://github.com/harry0703/MoneyPrinterTurbo.git
cd MoneyPrinterTurbo
uv python install 3.11
uv sync --frozen
cp config.example.toml config.toml

On Windows, use copy config.example.toml config.toml instead of cp. If you skip that line, the project creates config.toml by itself on first launch.

Get a free local AI model

Pull a model with Ollama (any model from ollama.com/library works; a small one is faster on a laptop):

ollama pull llama3.2
ollama list

Then open config.toml and set these lines (keep the model name exactly as ollama list shows it):

llm_provider = "ollama"
ollama_base_url = ""
ollama_model_name = "llama3.2"

Leaving ollama_base_url empty uses the local default, as the config file explains.

Make your first video (command line)

Put 2 or 3 of your own vertical clips in the project folder, then run:

uv run python cli.py --video-subject "black holes" \
  --video-language en-US \
  --video-source local --video-materials "./clip1.mp4,./clip2.mp4" \
  --video-aspect 9:16 \
  --voice-name en-US-AndrewNeural-Male

No clips of your own? Get a free key at pexels.com/api, put it in config.toml as pexels_api_keys = ["YOUR_KEY"], and drop the --video-source and --video-materials lines (Pexels is the default source).

Want 3 videos at once? See uv run python cli.py --help for the number-of-videos option and batch files. All options are listed there.

Prefer clicking? Use the WebUI

# Mac / Linux
sh webui.sh

# Windows
.\webui.bat

It opens in your browser. Type a topic, choose the voice and the clip source, click generate. API keys (if you use any) go in the basic settings.

Before you post

• Watch the whole video: AI scripts can state wrong facts. Fix the script with --video-script if needed.
• Use clips you have the right to use (your own, or Pexels/Pixabay under their licence).
• Add a personal touch (your face, your voice, your opinion). Fully automatic videos all look the same.