The support of multilingual content becomes crucial for educational platforms due to the benefits it offers. In this paper we propose a concept that allows content authors to use the power of the crowd to create (semantically) structured multilingual educational content out of their material. To enable the collaboration of the crowd, we expand our previously developed CrowdLearn concept and WikiApp data model to support multilingual educational content. The expanded concept, CoSMEC, was evaluated with an example implementation within the web-based educational platform SlideWiki. Based on this experience, we provide solutions for the most complicated technical issues we faced during the implementation. This paper also discusses statistics which show the flow of multilingual content usage. [This paper was presented at the OpenCourseWare Consortium Global Conference (Ljubljana, Slovenia, April 23-25, 2014).]