ZigZag: Free Embroidery Editor Built with AI to Escape $300 Machine Subscription

Background and Motivation

Developer Kathy Korevec created ZigZag after discovering that her Brother Skitch PP1 embroidery machine, priced around $300, required a paid subscription service to operate. Unwilling to accept the lock-in, she built her own free, open-source editor. Notably, she states the entire project was written with the assistance of AI — specifically ChatGPT.

Core Features

Input formats: Imports SVG vector graphics and converts them to machine embroidery stitches; also traces PNG and JPEG images into embroidery paths.

Stitch types: Supports 11 stitch types including fill, satin, and appliqué.

Export formats: Designs can be saved as PES, DST, EXP, JEF, or VP3 files.

Machine communication: Sends designs directly to the Skitch embroidery machine over Bluetooth.

AI Integration via MCP

ZigZag includes a built-in Model Context Protocol (MCP) server. This allows AI agents to draw and modify designs directly inside the current editing document. The author highlights this as unusual: most software treats AI as an auxiliary add-on, whereas ZigZag makes it a core design element.

Platform Support and Testing

Operating systems: Native binaries for Windows and macOS; Linux users can compile from Rust source using instructions in the README.

Test coverage: The codebase includes 136 library tests and 153 editor/MCP tests.

Known limitation: Bluetooth sending has not been verified on a physical Skitch machine, so that functionality may still have issues.

Existing Open-Source Alternatives

Before ZigZag, other open-source embroidery tools existed. Ink/Stitch is an Inkscape extension for embroidery design. Respira already enables sending designs to the Skitch without using Brother's proprietary Artspira software. ZigZag differentiates itself by combining a standalone editor, MCP-based AI integration, and a Rust codebase built entirely with AI assistance.

Project Website

Project website: https://zigzag.diy/. According to a tweet referenced by the author, the website itself was also constructed using ChatGPT.

}