Imagine telling your computer, "Design me a truss," and it generates a blueprint. That's the tantalizing promise of using Large Language Models (LLMs) for design scripting. A new research project explores how LLMs could bridge the gap between human creativity and the rigid logic of architectural design software. Traditionally, architects use visual scripting tools like Grasshopper to create complex geometric designs, but these tools require a steep learning curve and a shift to algorithmic thinking. This research proposes using LLMs as mediators, translating natural language prompts into the step-by-step instructions these tools understand. The prototype system works in layers: first, the LLM interprets the user's intent (e.g., "design a bridge"); then, it defines the necessary parameters (length, height, etc.); and finally, it generates a script that can be executed in Grasshopper. The results are promising, with the LLM successfully generating scripts for relatively simple structures like trusses and umbrellas. However, more complex designs, like suspension bridges, still pose a challenge. The LLM sometimes struggles to translate its abstract understanding into the precise syntax required by the software. While still in its early stages, this research offers a glimpse into a future where AI could make complex design tools more accessible and intuitive. Imagine architects conversing with AI to refine designs, explore variations, and push the boundaries of architectural innovation. This could democratize advanced design techniques and unleash a new wave of creativity. However, challenges remain, particularly in ensuring the reliability and originality of AI-generated designs. Future research will focus on improving the accuracy of the LLM’s output, exploring conversational interfaces, and incorporating multimodal inputs, bringing us closer to a world where designing buildings with AI is not just a dream, but a reality.