karasu weekly update: language v2.0, a focus canvas, and 97% fewer SVG elements
karasu is a text modeling language for system architecture, inspired by the C4 model. You describe your system in .krs files and karasu draws the views: system, deployment, and team structure. I started it about six months ago to learn agentic coding, and I haven't written its code by hand. Everything goes through Claude Code. This is the first of a weekly series. Each post covers what shipped…
This week on Karasu Weekly, the text modeling language for system architecture introduced a focus canvas to manage dense diagrams and released version 2.0 of its language, .krs. The focus canvas draws only the labels it can fit, while the full text resides in a separate canvas that opens upon clicking an edge. Notably, 97% fewer SVG elements were generated, cutting output size from 8.72 MB to 6.48 MB.
The .krs language v2.0 saw a closed vocabulary, with tags and annotations now belonging to the tool. Core notation elements like facet and boundary were formalized, and certain warnings were changed to errors. Smaller improvements included handling Unicode characters correctly and implementing edge bundling in gutters. Building with agents proved challenging, as a Claude Code PostToolUse hook accidentally executed any executable file an agent had edited, rather than formatting .ts files.
The workflow now includes a code review before PR creation to prevent such issues. Ten ADRs were recorded, and the design-doc-to-ADR loop continued. Next week's updates will include ordering channel lanes by edge turns and ensuring node and edge identity keys are collision-free.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
