Text-to-Lottie lets coding agents generate animations from text prompts

diffusionstudio/lottie is an open-source framework published under the MIT license that bridges natural language prompting and Lottie animation production. It works by installing a skill into a compatible coding agent — Claude Code, Codex, or any agent supporting the skills interface — using the command `npx skills add diffusionstudio/lottie`. Once installed, users describe the animation they want in plain text, and the agent handles workspace setup and file generation.

The repository has 3.1k stars and 162 forks at the time of publication.

The framework organizes output into scenes within projects. Scenes load automatically from `public/projects/` and are served through an included player that live-updates as the agent edits the underlying `lottie.json` files, allowing users to scrub and refine animations in real time. The project's prompt guide advises grounding prompts in concrete assets like SVGs or screenshots, using motion design terminology such as ease-in and ease-out, and thinking in camera-operator terms — describing pushes, pans, and zooms that the agent can simulate through group transforms. The repository has 3.1k stars and 162 forks at the time of publication.

Key facts

1. 01Open-source framework (MIT license) hosted at diffusionstudio/lottie on GitHub
2. 02Supports Claude Code, Codex, and any coding agent that supports the skills interface
3. 03Installed with `npx skills add diffusionstudio/lottie`
4. 04Animations are organized as scenes inside projects, stored in `public/projects/`
5. 05A built-in player live-updates as the agent edits `lottie.json` files
6. 06Prompt guide recommends grounding prompts in SVGs or screenshots and using motion design terminology
7. 07Repository has 3.1k stars and 162 forks on GitHub

Topics