alexgreensh/anidoodle

What it solves

Anidoodle allows users to create high-quality, hand-drawn style illustrations, animations, and music entirely through code. It replaces the unpredictability and recurring compute costs of AI image models with deterministic, code-based rendering, ensuring that the same source code always produces the identical image or film at any scale.

How it works

The project uses a core engine that treats every mark as a function. It provides 32 distinct artistic styles (such as watercolor, risograph, and pixel art) and a music engine that composes original scores from 21 different genres using data-driven notes rather than recordings.

Users can interact with it via AI agents (like Claude Code, Codex, or Grok Build) using plain-language requests, or drive the engine manually via Node.js tools. The system can output various formats including PNG, MP4, GIF, and self-contained HTML files that redraw the art from scratch upon opening.

Who it’s for

• Web Designers and Developers: For creating consistent hero illustrations, interactive mascots that react to user input, and branded assets.
• Marketers: For producing product launch films and social media content cut to a beat.
• Educators: For generating step-by-step drawing lessons with captioned timelapses.
• Brand Managers: For maintaining a consistent artistic "hand" across all company assets.

Highlights

• 32 Artistic Styles: Includes diverse mediums like sumi-e, charcoal erasure, and toy bricks.
• Deterministic Rendering: Guaranteed identical output across different machines and sizes.
• Code-Composed Music: Original scores generated from genre vocabularies with optional high-quality acoustic instrument samples.
• Interactive Web Art: Ability to create <ani-doodle> elements for websites that respond to cursors and forms.
• Lauch Film Kit: A specialized workflow to generate product launch videos with synchronized sound and motion blur.
• Style Matching: Can study an existing image to borrow its artistic hand for new subjects.