Imagine effortlessly reshaping a 3D model just by describing how you want it to change. That's the promise of ParSEL, a groundbreaking research project that lets you edit shapes using natural language. Traditionally, 3D modeling has been a technically demanding process, requiring specialized software and expertise. But what if you could simply tell the computer, "Make the chair wider," or "Extend the roof radially"? ParSEL makes this possible by translating your language instructions into precise, parameterized edits. This innovative system works by first breaking down a 3D model into its constituent parts, like the legs and seat of a chair. Then, it uses the power of large language models (LLMs), the same technology behind AI chatbots, to understand your edit request. However, LLMs alone aren't enough. While they excel at grasping the general idea, they often struggle with the specifics of geometric reasoning. For example, if you widen a chair's seat, the legs also need to be adjusted to maintain the chair's structure. To overcome this, ParSEL incorporates a clever algorithm called Analytical Edit Propagation (AEP). AEP uses mathematical equations to figure out how each part of the model needs to change to maintain consistency and structural integrity. This combination of LLMs and AEP allows ParSEL to handle complex edits while preserving the overall shape and relationships between its parts. The result is a powerful and intuitive tool that could revolutionize 3D modeling, making it accessible to a much wider audience. While still in its early stages, ParSEL offers a glimpse into the future of 3D design, where language becomes the primary tool for shaping the digital world. The research team is already exploring exciting applications, such as automatically generating variations of a shape based on a single request, opening up new possibilities for creative exploration and design automation.