Triangula is a generative art tool that converts images into high-quality triangulated and polygonal artwork using a modified genetic algorithm.
The tool solves the problem of creating stylized geometric representations of images through an iterative evolutionary approach. It works by applying a genetic algorithm that progressively refines a population of triangle or polygon configurations, evaluating each candidate against the source image and selecting the best-fitting variants for the next generation. The algorithm is optimized for images smaller than 3000 pixels with fewer than 3000 points, typically converging to a visually optimal result within a couple of minutes on desktop hardware, though a browser-based version is available for experimentation at reduced speed.
Triangula suits artists and developers who want to generate geometric abstractions from photographs or images without manual intervention. The project provides both a graphical interface and a command-line tool, making it accessible whether you prefer interactive parameter tuning or scripted batch processing. For most use cases, adjusting only the point count while leaving other parameters at their defaults produces good results, though the tool exposes fine-grained control over mutation behavior, population size, and rendering options for those wanting deeper customization. The desktop application runs significantly faster than the browser version, making it the practical choice for production work.
Development activity shows consistent maintenance with passing test coverage and active documentation. The project maintains separate repositories for its GUI and CLI implementations, indicating a deliberate architectural separation between interface layers and core algorithm logic.
