An open source AI video generator is one of two things: a tool that turns a topic into a finished short, or a model that turns a prompt into a few seconds of footage. This list covers both, with GitHub stars, licenses and real demos.
Full disclosure: we run MoneyPrinterTurbo Online, a hosted studio for the first kind. It does not run the other projects listed here.
These six open source AI video generator projects do the same core job: you type a topic and get a narrated short video. Each card opens a page on that project, with its own demos and setup steps.
Stars and dates are from GitHub on November 11, 2026. “Last update” is the last code push; check it before you install, because old projects break when the APIs they call change.
| Project | Stars | License | Last update | What it makes |
|---|---|---|---|---|
| MoneyPrinterTurbo (harry0703) | 142,000 | MIT | November 2026 | Topic to short with stock footage, voiceover, subtitles and music; web UI and API |
| MoneyPrinterV2 (FujiwaraChoki) | 32,400 | AGPL-3.0 | September 2026 | YouTube Shorts automation, plus a Twitter bot, affiliate marketing and outreach |
| Pixelle-Video (ATH-MaaS) | 30,000 | Apache-2.0 | June 2026 | Topic to short with an AI-generated image or clip for each sentence |
| MoneyPrinter (FujiwaraChoki) | 14,200 | MIT | March 2026 | The original: a topic becomes a YouTube Short |
| ShortGPT (RayVentura) | 8,200 | MIT | February 2025 | Shorts and TikTok framework with sourced footage, captions and dubbing |
| MoneyPrinterPlus (ddean2009) | 7,300 | GPL-3.0 | March 2025 | Batch shorts and clip remixing, with auto-publishing to Chinese platforms |
Four real outputs: three published by the projects’ own authors, one made in our online studio. Pixelle-Video draws an AI picture per sentence; the others cut stock footage to the script.
Pixelle-Video and MoneyPrinterPlus clips are from each project’s GitHub README.
The other kind of open source AI video generator. A model generates the footage itself from a text prompt or an image. It does not write a script, record a voice or add subtitles, and it needs a capable GPU.
| Model | Maker | Stars | License | Last update |
|---|---|---|---|---|
| Wan 2.2 | Wan-Video | 18,500 | Apache-2.0 | September 2026 |
| Wan 2.1 | Wan-Video | 17,400 | Apache-2.0 | March 2026 |
| CogVideoX | zai-org | 13,200 | Apache-2.0 | November 2025 |
| HunyuanVideo | Tencent Hunyuan | 12,800 | Its own license; read it before commercial use | October 2026 |
| LTX-Video and LTX-2 | Lightricks | 11,400 and 10,200 | Apache-2.0 | November 2026 |
| Mochi | Genmo | 3,800 | Apache-2.0 | October 2026 |
Some topic-to-video tools can call these models for you: Pixelle-Video lists Wan among the video models it supports. MoneyPrinterTurbo Online does not run or host any of them.
These four are often listed next to the tools above, but solve a different problem.
| Project | Stars | License | Last update | What it is for |
|---|---|---|---|---|
| RedditVideoMakerBot (elebumm) | 12,600 | GPL-3.0 | October 2026 | Turns Reddit threads into narrated videos with one command |
| NarratoAI (linyqh) | 11,800 | MIT | October 2026 | Writes a narration for a video you already have and edits it to match |
| story-flicks (alecm20) | 2,600 | None stated | March 2025 | Story videos with AI-generated illustrations |
| short-video-maker (gyoridavid) | 1,500 | MIT | June 2025 | Short videos for TikTok, Reels and Shorts, driven through the Model Context Protocol |
It depends on the job. For finished narrated shorts, MoneyPrinterTurbo is the most starred and most actively updated open source AI video generator. For AI-illustrated shorts, Pixelle-Video. For original footage from a prompt, the Wan and LTX models are the most active.
MoneyPrinterTurbo: broad topics, stock footage, a web UI and regular releases.
Pixelle-Video, if you can run ComfyUI or pay for image APIs.
MoneyPrinterV2 for YouTube and Twitter, MoneyPrinterPlus for Chinese platforms.
A Wan or LTX model, if you have a GPU and will add script, voice and captions yourself.
Yes. Every open source AI video generator on this page is free to download from GitHub. Running one is not always free: most call paid AI APIs for the script, voice or images, and video models need a GPU. MoneyPrinterTurbo can run at no cost with a local model, free stock footage keys and the free Edge voices.
Check the license if you plan to build a product on top. MIT and Apache-2.0 are permissive; GPL-3.0 and AGPL-3.0 require you to share your changes under the same terms.
None of the six tools above has an official hosted version; they are code you run yourself. Our studio does their core job in the browser, on the MoneyPrinterTurbo workflow. Real users made these eight from one line of text. Tap one to play it with sound.
Videos by MoneyPrinterTurbo Online users. Footage from Pexels.
Open the studio, type a topic, pick a voice, a size and a caption style, then generate and download the MP4. It takes a few minutes, a quick way to see a result before installing an open source AI video generator. The full guide is how to use MoneyPrinterTurbo.
Sign in with Google or GitHub and describe the video in a sentence, or paste your own script to have it read word for word.
Choose from 17 languages, a target length, 9:16, 16:9 or 1:1, a voice, a caption style and the music.
The button shows the credit cost first; a one-minute video costs 48 credits and new accounts get 60. Then download the MP4.
Not included online: AI-generated images or clips for each scene, automatic posting to YouTube or TikTok, video translation and dubbing, and your own choice of models. For those, install the project that does it.
A new account gets 60 credits, enough for about a one-minute video. Paid plans start at $19 a month, and every plan and credit pack includes priority generation.
Every plan includes
Yes, many. This page lists 15: six tools that turn a topic into a finished short (MoneyPrinterTurbo, MoneyPrinterV2, Pixelle-Video, MoneyPrinter, ShortGPT, MoneyPrinterPlus), four tools for other jobs, and the Wan, CogVideoX, HunyuanVideo, LTX and Mochi video models. All are on GitHub.
For topic-to-video tools, usually yes: MoneyPrinterTurbo asks for 4 CPU cores and 4 GB of RAM and treats a GPU as optional, because the heavy work goes to APIs. Video models such as Wan, LTX and HunyuanVideo are built for GPUs and are slow or unusable without one.
Every open source AI video generator here can be self-hosted; that is how they are meant to be used. Topic-to-video tools usually run with Python or Docker on an ordinary computer and call APIs for the heavy work. Video models need a GPU of your own or a rented cloud GPU.
No. MoneyPrinterTurbo Online is an independent hosted service built on the open-source MoneyPrinterTurbo workflow, with its own accounts and credits. The open source AI video generator itself is the GitHub repository by harry0703, which you can install and run for free.
Type a topic and get a finished short in your browser, with no install and no API keys. The first video is on us.
