Abstract
The article presents the development of an intelligent system capable of automatically generating web content based on a user's text query using a language model built on the GPT architecture. It discusses the objectives, methodology, and stages of designing and implementing a web application that synthesises the structure and visual design of websites based on a description given in natural language. A key component of the system is the integration of a neural network model via the OpenAI API. The architecture of the software solution, the principles of interaction between its modules, and the sequence of stages for generating and subsequently verifying HTML/CSS code are described. A comprehensive assessment of the functionality, performance, and practical applicability of the system has been carried out. The results of the experimental study demonstrate the high reliability, scalability, and potential for application of the developed tool in educational, professional, and commercial spheres.