This paper introduces a novel methodology for generating data designed for cluster analysis. The general concept of the generator originates in model-based clustering methods and their cluster definition. The newly proposed methodology broadens the options for tailoring the characteristics of the resulting dataset compared to existing methodologies and the customization options they provide. The most significant advancements in generating datasets with clusters include the capability to produce datasets with various variable types and different cluster shapes. The approach combines the NORTA algorithm with Cholesky decomposition, enabling users to customize the resulting dataset to their preferences precisely. This generator makes generating datasets with desired variable types and clusters of differing quantities, shapes, and separations remarkably straightforward.