As information systems are dealing with a growing amount of unstructured text information, text mining technologies and their corresponding software tools have become increasingly widespread. However, text mining is still a young discipline that is starting to witness the advent of flexible, robust, and general software architectures that leverage the abundance of research results in benefiting application developers. In this article, we introduce Plato, a new software architecture that is unique in the set of text mining functionalities it provides to general purpose applications. This system is described and compared to other classification, information retrieval, and information extraction systems. We also include two case studies of applications developed with Plato and currently in use.