MoneyPrinterTurbo

// Overview

MoneyPrinterTurbo is a self-hosted short-video generator. You give it a topic or a keyword and it runs the whole pipeline: an LLM writes the script and extracts the search keywords that decide the footage, clips are pulled from stock libraries or generated, a text-to-speech voice reads the narration, subtitles and background music are added, and the result is rendered as an HD vertical, landscape or square video. Every stage stays adjustable — you can supply your own script, your own clips, or your own audio instead.

It runs four ways: a Streamlit WebUI, an HTTP API with OpenAPI docs, a CLI, and an agent workflow in which a coding agent reads the project's SKILL.md, installs and configures everything, and hands back the video file path. Deployment is Docker Compose against a prebuilt GHCR image, or a manual Python 3.11+ setup with uv; Windows users have a one-click package and there is a Google Colab notebook for trying it without a local install.

Model choice is deliberately open. Scripts can come from Kimi, OpenAI, Anthropic Claude, Gemini, DeepSeek, Qwen, Azure OpenAI, Volcengine Ark, xAI Grok, MiniMax or Xiaomi MiMo, or through gateways and local runtimes such as OpenRouter, LiteLLM, Ollama and OneAPI. Footage can be your own files, free stock from Pexels, Pixabay or Coverr, or generated by a text-to-video service. Voices include Edge TTS, which needs no API key at all. Finished videos can be published straight to TikTok, Instagram and YouTube Shorts.

// What it does

• Topic or keyword to finished video: script, footage matching, voiceover, subtitles, background music and editing, with control retained at every stage
• Four ways to drive it — WebUI, HTTP API, CLI, or an AI agent that reads the project's SKILL.md and sets itself up
• Script models from Kimi, OpenAI, Claude, Gemini, DeepSeek, Qwen, Azure OpenAI, Volcengine Ark, Grok, MiniMax and Xiaomi MiMo, or any OpenAI-compatible gateway or local runtime
• Footage from your own uploads, free stock APIs (Pexels, Pixabay, Coverr), or text-to-video services, with adjustable clip duration and ordering
• Voiceover through Edge TTS (free, no API key), Azure Speech, Gemini, MiniMax, ElevenLabs, Fish Audio and others, with previewable samples
• Batch generation of output variants, task history, importable and exportable settings, and 9:16 / 16:9 / 1:1 export with direct publishing to TikTok, Instagram and YouTube Shorts

// Getting started

Docker Compose is the quickest path on macOS and Linux; a manual uv-based install works too. Python 3.11 or later is required for local deployment, and a GPU is optional (it mainly helps local transcription and batch rendering).

git clone https://github.com/harry0703/MoneyPrinterTurbo.git

cd MoneyPrinterTurbo
docker compose -f docker-compose.release.yml up

cd MoneyPrinterTurbo
uv python install 3.11
uv sync --frozen
sh webui.sh

# WebUI: http://127.0.0.1:8501
# API docs: http://127.0.0.1:8080/docs

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

// When to use it

• Produce a batch of vertical shorts on a topic without touching a video editor
• Keep the whole pipeline self-hosted and swap in whichever LLM, TTS and stock-footage providers you already pay for
• Generate narrated explainer clips in several languages from one script, using free Edge TTS voices
• Drive video generation from your own backend through the HTTP API rather than the WebUI

// How MoneyPrinterTurbo compares

MoneyPrinterTurbo alongside other open-source video generation tools AI/TLDR tracks, ranked by GitHub stars.