Show HN: Open-sourcing our text-to-CAD app (github.com)

A new open-source app called CADAM has been released, enabling users to generate parametric 3D CAD models directly from natural language descriptions and images. Running entirely in the browser via WebAssembly, CADAM leverages AI—specifically the Anthropic Claude API—to interpret user input, automatically extracting adjustable parameters that users can tweak instantly through interactive sliders. This real-time preview functionality uses Three.js and React Three Fiber for smooth model visualization, supporting exports in popular CAD formats like .STL and .SCAD. The significance of CADAM lies in its seamless integration of AI-driven language understanding with traditional CAD modeling, democratizing 3D design by making it accessible without deep technical expertise. With its modular architecture—built on React, TypeScript, Supabase backend, and open source libraries like BOSL, BOSL2, and MCAD—the project invites community contributions to further its capabilities. Developers can run the app locally with detailed setup involving Node.js, Supabase, and ngrok for webhook handling, ensuring easy experimentation and extension. Overall, CADAM exemplifies a growing trend in AI-assisted creative tools, combining generative language models with parametric design to accelerate prototyping workflows in engineering and maker communities.