Prompt-to-Geometry

Exploring Large Language Models for Generative Python Scripting in Architectural Design

Prompt-to-Geometry examines how large language models can support reusable Python scripting and generative architectural design workflows. The research translates natural-language design instructions into code-driven geometric studies, linking prompt development, script iteration, and three-dimensional modeling.

The undergraduate research project compares generated design variations with parametric models and visualizes the relationship between text prompts, computational definitions, and architectural form.