Planning, scripting, recording narration, designing subtitles, editing. A single short typically takes 2-3 hours, right? Now you just type a topic in the terminal and you're done. AI handles everything from script to final video.
Gemini free tier + Edge-TTS (free) + FFmpeg (free). Completely free.
What Is It?
The short-form video market is booming. Valued at roughly $34.8 billion in 2024, it's projected to grow at over 30% CAGR to reach $289.5 billion by 2032. Over 90% of marketers report positive ROI from short-form video ads.
But here's the problem — consistently creating short-form videos is genuinely exhausting. You need to post 3-5 times a week to ride the algorithm, and planning, filming, and editing every single one is nearly impossible for individuals or small teams.
Open-source AI shorts generators are popping up to solve exactly this. The core technology boils down to three things:
Models like Gemini, GPT-4, and DeepSeek automatically write video scripts from just a topic.
Microsoft Edge-TTS provides 300+ high-quality voices without an API key. Gemini 2.5's native TTS can even handle emotional expression.
FFmpeg's Ken Burns effect (that slow zoom-in/zoom-out you've seen) turns static images into dynamic video.
In February 2026, @build_daemon shared an "AI Shorts Auto Generator" on Threads that went viral with 328 likes. Similar tools already exist on GitHub: MoneyPrinterTurbo (49,500 Stars), ShortGPT (7,100 Stars), and the MCP-integrated Short Video Maker.
What's Different?
Sure, there are paid services like Runway, Pika, and HeyGen. But the open-source tools we're looking at are a different breed entirely.
| Paid SaaS (Runway, Pika, etc.) | Open-Source Generators | |
|---|---|---|
| Cost | $8–95/month (by plan) | Free (only minor API costs) |
| Control | Limited to platform templates | Full code-level customization |
| Video Style | AI-generated video (live-action/animation) | Slides + narration + subtitles (info-driven) |
| Mass Production | Credit limits apply | Unlimited (runs locally) |
| Best For | Ads, music videos, visual effects | Educational, news, summary content |
| Technical Difficulty | A few clicks in a browser | Requires Python & terminal basics |
In short: Runway and Pika are great for making "polished-looking videos", while open-source tools excel at producing "consistent daily content".
If you're a channel operator who needs to post shorts daily, a marketer repurposing blog content into video, or a creator mass-producing news summaries — open-source tools are the clear winner.
Let's compare the major tools. For beginners, we recommend MoneyPrinterTurbo.
| Project | Stars | LLM Support | TTS | Key Features |
|---|---|---|---|---|
| MoneyPrinterTurbo | 49.5k | GPT, Gemini, DeepSeek, Qwen + 12 more | Edge-TTS, Azure | Web UI, batch generation, largest community |
| ShortGPT | 7.1k | OpenAI | ElevenLabs, Edge-TTS | 30 languages, built-in translation engine |
| Short Video Maker | 965 | MCP integration (any LLM) | Kokoro TTS | MCP/REST API, Docker deploy, video in 30 seconds |
| @build_daemon | New | Gemini | Free TTS | Ken Burns effect, one-click automation, cross-platform |
Quick Start Guide
We'll walk through MoneyPrinterTurbo since it has the largest community. The flow is similar for other tools.
Install Python 3.10+, FFmpeg, and ImageMagick. On Mac: brew install ffmpeg imagemagick. On Windows, download from their official sites.
git clone https://github.com/harry0703/MoneyPrinterTurbo.git
cd MoneyPrinterTurbo
pip install -r requirements.txt
python webui.pyThe web UI opens in your browser.
Get a free Gemini API key from Google AI Studio → enter it in the web UI settings. Select Edge-TTS for completely free narration.
Be specific, like "Explain Bitcoin halving in 30 seconds." Set the aspect ratio to 9:16 for direct upload to Shorts/Reels.
If you prefer the CLI, check out @build_daemon's project. The cinematic Ken Burns effect is its standout feature.
Want to Go Deeper?
Core Tools
MoneyPrinterTurbo — GitHub Repository The undisputed leader with 49,500 Stars. Has a web UI so you can get started without touching code, and supports 12+ LLMs with Edge-TTS. https://github.com/harry0703/MoneyPrinterTurbo
ShortGPT — GitHub Repository The multilingual powerhouse supporting 30+ languages. Includes auto-subtitles, auto source video collection, and a built-in translation engine. https://github.com/RayVentura/ShortGPT
Short Video Maker — GitHub Repository A next-gen tool with MCP protocol support. Docker deployment, 30-second video in one minute. https://github.com/gyoridavid/short-video-maker
Technical References
Gemini API — TTS Documentation Guide to Gemini 2.5's native TTS capabilities. Supports 24 languages, emotional expression, and multi-speaker output. https://ai.google.dev/gemini-api/docs/speech-generation
Edge-TTS — GitHub Repository A Python package offering 300+ high-quality voices for free, no API key required. https://github.com/rany2/edge-tts
Bannerbear — FFmpeg Ken Burns Effect Guide Tutorial on implementing the Ken Burns effect using FFmpeg's zoompan filter. https://www.bannerbear.com/blog/how-to-do-a-ken-burns-style-effect-with-ffmpeg/
Pixazo — Top 10 Open-Source AI Video Generation Models in 2026 Comparison of the latest models including HunyuanVideo, CogVideoX, and SkyReels. https://www.pixazo.ai/blog/best-open-source-ai-video-generation-models
