Abstract
Prototyping complex computer-aided design (CAD) models in modern softwarescan be very time-consuming. This is due to the lack of intelligent systems thatcan quickly generate simpler intermediate parts. We propose Text2CAD, the firstAI framework for generating text-to-parametric CAD models usingdesigner-friendly instructions for all skill levels. Furthermore, we introducea data annotation pipeline for generating text prompts based on naturallanguage instructions for the DeepCAD dataset using Mistral and LLaVA-NeXT. Thedataset contains $\sim170$K models and $\sim660$K text annotations, fromabstract CAD descriptions (e.g., generate two concentric cylinders) to detailedspecifications (e.g., draw two circles with center $(x,y)$ and radius $r_{1}$,$r_{2}$, and extrude along the normal by $d$...). Within the Text2CADframework, we propose an end-to-end transformer-based auto-regressive networkto generate parametric CAD models from input texts. We evaluate the performanceof our model through a mixture of metrics, including visual quality, parametricprecision, and geometrical accuracy. Our proposed framework shows greatpotential in AI-aided design applications. Our source code and annotations willbe publicly available.