An open-source text-to-CAD tool that produces 3D models from text and image references by generating CAD as code.
An open-source text-to-CAD tool that produces 3D models from text and image references by generating CAD as code.
You describe a part in plain English, optionally with reference images. The app uses Claude to write OpenSCAD code, which runs in the browser through WebAssembly and shows a live Three.js preview. It pulls out adjustable dimensions as sliders, so you can change sizes without asking the AI to regenerate. You can export STL, SCAD or DXF.
The README's benchmark gallery shows single prompts producing parametric models, from knurled knobs and threaded bolts to a V8 engine with 22 dimension controls. It gives no failure rates or accuracy measurements.
Self-hosting needs Node, the Supabase CLI, ngrok for webhooks, and your own API keys (Anthropic and others). The hosted demo at adam.new/cadam needs no install.
Write a parametric CAD script in [CadQuery or OpenSCAD] for [describe the part, e.g. a wall bracket for a 25mm pipe], with named variables for all key dimensions at the top. Run it and export an [STL or STEP] file to this folder. Stop when the file exports without errors and show me the list of parameters I can change.
Score every demo request against your ICP, update HubSpot and ping the right rep in Slack within a minute of the form submit.
Every new Zendesk ticket gets a category, priority and sentiment tag plus a draft reply in an internal note before an agent opens it.
Copilot and Claude review each PR for bugs, missing tests and risky changes before a human looks, so reviewers can focus on design.