AI & text-to-CAD

LLM

Large language model

Definition

An LLM is a large neural network trained on text to predict and generate language. In CAD, LLMs are used to write modeling code, drive a CAD program's API, answer software questions, or translate a prompt into structured modeling steps.

Also known as: large language model, language model

Updated

How LLMs reach geometry#

An LLM does not produce a solid by itself. It produces text, and that text becomes geometry through code or tools. The GPT-3 paper showed that large models can do new tasks from a few examples in the prompt, which is why they can write OpenSCAD, CadQuery, or CAD API scripts without special training.

Protocols such as MCP connect a model to a running CAD program, so it can issue commands and read results. Accuracy still depends on the model's spatial reasoning, which is weaker than its fluency with code, so generated dimensions and feature choices need checking.

Sources

All glossary terms