Welcome to AutoShorts
AutoShorts is a fully fledged framework that generates shorts videos with the help of AI to write the script. It can generate popular types of video seen on YouTube Shorts and TikTok.
If you want to support the development of this package, consider buying me a coffee:
The package author is not responsible for any misuse of the package, any content generated by the package, and any loss arising from the use of the package. Use at your own risk. Package is subject to change and may have breaking changes in the future. Not meant for production usage.
Introduction
The package is built with a flexible abstraction layer which allows you to quickly create different types of videos with different AI tools and can be easily extended to include more types and tools internally.
This repo includes a CLI interface and JS interface. The CLI interface is built on top of the JS interface. The JS interface is the core of the package and can be used to generate videos programmatically.
This package is in the early stages of development and may have bugs - especially when interacting with AI to generate scripts. This package is not meant to be used in production environments yet. Since AI output is unpredictable, work is being done to make the output more predictable and controllable. If the AI generates an incorrect JSON output, then you can manually edit the JSON output to fix the issue. This package is subject to change and may have breaking changes in the future. Use at your own risk.
Features
You can generate popular types of short videos seen on YouTube Shorts and TikTok like:
- Text-to-Speech (TTS) videos
- Topic-based videos with images
- Text message story videos
- 'Would you rather' videos
- Ranking item videos
- And more...
The topic-based video type can be used to generate videos on a wide range of topics like:
- News-related videos
- Educational videos
- Fun facts videos
- And more...
| Text Message Video | News Video | Quiz Video | Rank Video | Would You Rather Video |
|---|---|---|---|---|
 |  |  |  |  |
You can request new video types and tools to be added to the package by creating an issue on the GitHub repo.
Installation
Note: By default, the package use Ollama to generate scripts. Therefore a working Ollama installation is required. It is recommended to use the llama3.2 model for best results.
You can install this model by running the following command: ollama pull llama3.2.
Using other local models may result in incorrect output.
You can use other AI APIs such as OpenAI ChatGPT 4o, Google Gemini AI, and Anthropic Claude by providing the necessary API keys and setting the aiType parameter to the appropriate value.
Ultimate Goal
The goal of this package is move away from short video generation to a more general video generation framework that can generate videos of any length and type. This package is a stepping stone to that goal. If you have any ideas or suggestions, feel free to create an issue on the GitHub repo. Also any financial support is greatly appreciated.