lemo-opuscar

Lemo-Opuscar is an open-source project published on GitHub by lemomo-ai that has an AI coding agent make short films entirely in code. You supply the story and the agent directs: frames are drawn one by one with Canvas and WebGL, music comes from free sample libraries, and narration is generated with text-to-speech (TTS). The project says it uses no video generation models and no stock footage. It includes 43 styles, covering hand-drawn work, East Asian painting traditions, printmaking, cartoons, games and classic cinema, each with sample films. The author says every frame, note and cut in the samples was written in code by Claude Opus 5.5. Default output is 1920×1080 at 24 frames per second.

Who it's for:people who already use an AI coding assistant such as Claude Code and want to make animated shorts or explainer videos. Setup happens on the command line, so readers who don't write code will find it hard to get started.

Before you start: the recommended install is as a Claude Code plugin, or you can clone the repository and run it locally. It needs Node.js 20 or later, ffmpeg and Python 3.11 or later; the agent downloads and installs the remaining dependencies. On first use it downloads guides, tools and style assets to a lemo-opuscar folder in your home directory, and finished films are saved in a films folder in the current working directory. The project notes that one film takes about 30 to 60 minutes and uses a lot of tokens, billed by whichever AI service you use. Licensed under MIT; third-party fonts, music and sounds in the samples keep their own licenses, and you are responsible for any material you add to your films.

Based on the project's public documentation. Not tested by AI Inspiration World.