a design tool that thinks in two-by-fours

Someone planning a deck or a workbench doesn't want CAD. They want to know what it will cost and how to cut it. General 3D tools can draw anything, but they know nothing about lumber: not that a two-by-four is really an inch and a half by three and a half, not how far a joist can span, not that every saw cut costs an eighth of an inch. I wanted a design tool that started from the lumber aisle and ended with something you could build from.

StickBuild is a 3D editor in the browser where every board is a real size and moving one feels like moving a shape in Figma, with snapping, smart guides, and measurements you can type over. The idea that makes it more than a drawing tool is the parametric assembly. You drop in a deck and give it a size, and the tool frames it from a span table, adds a beam and posts when the span demands them, and staggers the deck boards so each joint lands on a joist. Walls, roofs, stairs, railings, and footings work the same way, and any of them can be broken apart into individual boards.

From that one list of parts the tool produces a shopping list packed into stock lengths and priced against current store listings, and a printable plan set with dimensioned elevations, a cutting diagram, and a step-by-step build guide. An AI assistant can add and change parts on request, with its actions checked against the same rules the editor enforces.

I built it alone, with Claude Code as a pair, over about eleven weeks. What made that workable was writing the project's engineering rules into the repository itself, so that any change which loosened a structural limit was stopped unless I had asked for it.

StickBuild is live and free in beta, with a public demo, a library of starter plans that anyone can copy and resize, and shareable view links. For me it was proof of how far one designer can take a real product with AI as the engineering partner, and it changed how I approach that work with clients: the craft is in deciding what must never silently change.