rayventura/shortgpt

ShortGPT is a Python framework that automates the process of creating short-form videos for platforms like YouTube Shorts and TikTok. Instead of manually writing scripts, recording voiceovers, finding footage, and editing clips together, you describe what you want and the framework handles the production steps automatically using AI. The framework has several distinct engines depending on what you are making. The shorts engine handles everything from generating a script to final video output, including adding the metadata needed to post on YouTube. The video engine handles longer content, automatically sourcing background footage, timing captions, and preparing audio. A translation engine takes an existing video or YouTube link, transcribes what is said, translates it into another language, generates a new voiceover, and outputs a completely dubbed video. Voiceover generation supports over 30 languages through Microsoft's EdgeTTS service, with ElevenLabs as an alternative for higher quality voice synthesis. Background visuals are sourced automatically from Pexels (a library of free stock footage and photos) and Bing Image search. Captions are generated and timed automatically. The underlying scripts and editing instructions are produced by OpenAI's language models, which also drive decisions about pacing and structure. Video processing is handled by a Python library called MoviePy. You can run ShortGPT locally using Docker, which packages all the dependencies so the setup is consistent regardless of your system. A Google Colab notebook is also available if you want to try it without installing anything on your own computer. Colab is a free cloud service that runs Python notebooks in a browser.