There are many tasks in the field of analysis of legal texts, which haven't been resolved fully nowadays. Creating solutions for some of these tasks, which applicable to business, requires the introduction of a full-fledged information system. In this article several stages of design and development of such system for search of semantically important connections between decisions of the constitutional court are considered. Describes how to create a vertical prototype and obtained result of processing of texts. Also describes the problems of scalability and organization of the database, proposed modifications to solve these problems and introduces way of the transition from the existing prototype to this modification.