text-to-cad
A library of agent skills that let an AI coding agent generate, inspect, source and slice CAD and robot-description files, from creating a part in STEP format to writing URDF robot structure files and checking manufacturability.
text-to-cad is a library of agent skills for computer-aided design, engineering and manufacturing, letting an AI agent generate, inspect, source, slice and hand off CAD and robot-description artifacts from local project files.
Its skills include creating and editing CAD models from plain-language or image requests with STEP, STL, 3MF or GLB output; previewing CAD and robot files in the browser; finding off-the-shelf STEP parts such as screws, bearings and motors; producing dimensioned engineering drawings and 2D DXF files; writing URDF robot structure files and adding MoveIt planning groups through SRDF; creating simulator worlds in SDF; checking files before upload to a cutting service; measuring mesh printability for additive manufacturing; reviewing a part for sheet metal, CNC or injection molding manufacturability with cited rules; and slicing supported mesh files to G-code.
It is written in Python and released under the MIT license.