What you need first
- A computer (Mac, Windows or Linux). The README asks for at least 4 CPU cores and 4 GB of RAM. A graphics card isn't required.
- Python 3.11 or newer (Mac and Linux). Open your terminal and type
python3 --version. If it's older or missing, install 3.11 from python.org. On Windows you can skip Python and use the one-click package (step 1 below). - Three things for the video: an AI model that writes the script, a stock-footage key for the clips, and a voice. Two of them are free with no card. The third has a free option too.
- About 20 minutes the first time.
The free keys, one by one
1. Clips: Pexels (free). Go to pexels.com/api, sign up and request your API key. Copy it. This is where the clips come from by default.
2. Clips, second source: Pixabay (free, optional). Go to pixabay.com/api/docs, log in, and your key is shown on that page. Add it if you want a second library of clips.
3. Voice: no key at all. In the app, the voice service called "Azure TTS V1" runs on Edge TTS, which is free and needs no key. Pick an English voice such as en-US-AriaNeural or en-US-AndrewNeural.
4. The script writer: pick one.
- Free, on your own computer: Ollama. Install it from ollama.com, then download a model in your terminal, for example
ollama pull llama3.2. In MoneyPrinterTurbo, choose Ollama as the provider and type the model name. It's free, but slower on an older computer, and small models write weaker scripts. - A key you already pay for. The app works with OpenAI, Anthropic Claude, Google Gemini, DeepSeek and others. Use a key from a service you already use. Each script costs a little on that provider's bill.
- Claude subscription. If you pay for Claude Pro or Max and have Claude Code installed and logged in on the same computer, the app can use it instead of an API key.
You don't need any of the paid sponsors listed in the README. The default provider in the settings file is one of them (Kimi), so change it to the one you picked.
The setup steps
Windows (the easy way)
- On the repo page (github.com/harry0703/MoneyPrinterTurbo), click Releases on the right, open the latest one and download the
.7zfile from Assets (not "Source code"). - Extract it, double-click
update.bat, then double-clickstart.bat. The app opens in your browser.
Mac or Linux
- Open your terminal and make a folder for it:
mkdir mpt && cd mpt - Create a safe space for it:
python3 -m venv .venv - Turn it on:
source .venv/bin/activate - Install the official package:
python -m pip install mpt - Start it:
mpt webui. The app opens in your browser.
Next time, repeat steps 3 and 5 from the same folder.
Then, in the app
- Click Settings (top right) and paste your Pexels key (and Pixabay, if you made one).
- Pick your script writer and paste its key, or choose Ollama and type the model name.
- Under Audio Settings, choose Azure TTS V1 and an English voice. Click Preview Voice to hear it.
Heads-up: the official package is called mpt. A package named "MoneyPrinterTurbo" on PyPI is made by someone else, so don't install that one.
Your first Short
- In Video Subject, type one word or a short topic, for example
coffee. - Under Video Settings, keep Portrait 9:16.
- For your first test, set Background Music Source to No Background Music or Custom Background Music with your own track (see the honest part).
- Click Generate Video and wait. A short video takes a few minutes, longer with Ollama.
- Your video is in the app's data folder, under
storage/tasks. On a Mac that's~/Library/Application Support/MoneyPrinterTurbo; on Windows it's%LOCALAPPDATA%\MoneyPrinterTurbo. - Watch it all the way through before you do anything with it.
The honest part
- Your keys stay on your computer. They sit in a settings file (
config.toml) in the app's folder. Don't share that file, don't show it on a screen recording, and be careful with the app's "export settings" option, because it can include your keys. - Read the script before you post. The AI can get facts wrong. Fix it in the script box, or paste your own script.
- The clips are generic stock footage. Good for a test, but for anything that matters, upload your own clips (the app accepts local images and videos).
- The music. The README says some of the built-in background songs come from YouTube videos and asks you to delete them if there's a copyright problem. Use No Background Music or a track you have the rights to.
- Say it's an AI voice. If you post a video with a generated voice, label it on the platform and say so in the caption.
- The app can post for you to TikTok, Instagram and YouTube. Leave that off until you've watched and fixed the video yourself.
- It changes often. If something fails, update it:
python -m pip install --upgrade mpt(Mac/Linux) orupdate.bat(Windows).
As of October 22, 2026, MoneyPrinterTurbo has over 133,000 stars on GitHub and its latest release is v1.3.9 (October 21, 2026). Menus, versions and providers change.
Sources: github.com/harry0703/MoneyPrinterTurbo (README-en.md, config.example.toml, docs/voice-list.txt, Releases), pypi.org/project/mpt, read October 22, 2026.