ClassCAD is the engine. Buerli, the SDK for your own CAD applications, costs nothing. ClassCAD.ai, with which AI agents model and write code, is open source. Buerligons, the finished CAD application, is free to use on Free. Money changes hands in one place only: the ClassCAD engine, through a plan or a partnership.
Our customers meet four websites: classcad.ch, classcad.ai, buerli.io and buerligons.io. This post explains how they belong together.
ClassCAD is the CAD engine. ClassCAD.ai is the tool with which AI agents model on it and write code. Buerli is the free SDK for your own applications, and Buerligons the finished CAD application. ClassCAD PDM for versioning and release is coming. You pay only for the engine, never for the SDK, the AI tools or Buerligons on Free.
In August we described four AI projects around ClassCAD. Two months later the picture has changed enough to write it again. The pieces now live in one open-source monorepo, classcad-ai, as five npm packages:
@classcad/skill — the verified knowledge of the API@classcad/script — the medium: an agent writes a program, not one call per turn@classcad/renderer — the proof: deterministic views, sections and technical drawings@classcad/mcp — the MCP server for coding agents, now self-contained, with its own engine@buerli.io/ai — the in-app assistant for end usersAnd the part we are most pleased with: every MCP session comes with a live CAD app. Buerligons docks into the agent's session, so the user turns, selects and edits the same model the agent is building, and a share link lets a colleague do the same from another machine.
Our custom piping CAD application demonstrates how to build a minimal yet powerful CAD solution on top of our WASM-based CAD engine. The full source code is available in our starter repository:
👉 https://github.com/awv-informatik/buerli-starter
This article is for companies who need CAD software that standard tools cannot provide — and therefore decide to build their own.
Custom CAD is often misunderstood as a collection of UI tweaks, scripts, or isolated automation features. In reality, custom CAD is a layered system that transforms CAD from a general-purpose design tool into a domain-specific, data-driven model-building platform.
Below is a clear framework describing the five core dimensions of custom CAD, from user interaction down to automated model generation.
The landscape of programmable CAD is evolving rapidly. From browser-based platforms to headless geometry engines, developers now have a range of tools to express and automate CAD logic through code.
Code is CAD’s native language for machines.
For decades, CAD has been optimized for humans clicking, dragging, and sketching. That made sense when design lived on individual workstations and geometry was authored by hand. But machines don’t think in sketches. They think in code.
If CAD is going to be machine-producible, machine-editable, and generative-ready, then code isn’t a convenience — it’s a necessity.
We extend the feature list by adding :
Extrude along path, pipe sweep and other. These features will be available with the next release in Part and in Solid API.
The glowbuzzer React library allows customers to easily create a sophisticated 3D scene of their robot or machine environment using three.js. However, many customers need to import STEP files into the scene and then want to be able to interact with the geometry of the model to pick out points, lines and so on. These are used in the resulting program (set of moves) transmitted to the machine or robot.
We have created an online configuration tool for one of our products and wanted to provide an automatically generated 3d model and drawings to our customers. The customers should be able to get the correct models based on their input values almost instantly, without the involvement of an engineering department.
The company manufactures components for electronic component testing at the end of series production. Especially in the automotive sector, due to the multitude of different connectors, several new test adaptations are required for each new product. The high degree of customization of the products produced in small series results in high development effort, high costs, and long lead times. The startup has set itself the goal of significantly reducing development effort through digitization and automation, thereby becoming fast and economically attractive.