Problem of safety ensuring of children and underage adolescents in Web–space is becoming increasingly acute. Younger generation information security is the state of children and adolescents protection, in which there is no risk associated with harm to their health and (or) physical, mental, spiritual, moral development from destructive information. This article describes the development of new software product that allows analyzing text from files, Web sites, pages, applications, etc. Analog programs were reviewed; modern text recognition algorithms were described. Screenshots and flowchart of software parts designed functioning for semantic text analysis in Web space are presented. Any Web page is file with extension, and after analyzing, it is possible to evaluate the information and classify it by topics, as well as to determine whether there is destructive content. In the presented study, the software tool was developed for semantic analysis and formalization of text. Exact trigger points from 26 to 30 times evaluate the algorithm functioning. With accuracy of up to 80%, it can be argued that algorithm is suitable for determining the orientation of article and content of ideas expressed in it, taking into account duality of model.