One of the new educational technologies that has shown its undoubted effectiveness is e-learning. In developed countries, e-learning covers all levels of education and is widely used not only in universities, but also in high school and in the organization of corporate (postgraduate) education [1]. Such platforms require the elaboration of high-quality and relevant teaching resources, the constant updating of existing ones. This, in turn, is a complex process consisting of processing a variety of materials, their analysis, synthesis, creative development and processing of all elements to build a single harmonious structure [2]. Up to now, far too little attention has been paid to dynamic content generation for e-learning courses. The aim of this work is to describe development steps of the application that generates the content from the user request (key words). The program is based on our model for creating dynamic content of online training courses in Romanian, English and Russian languages. According to our approach of the program model for the dynamic creation of training courses, we have 6 phases. At step 1 we managed to implement phase 1 and develop a web-crawler that creates networks of synonymous words. In this article we describe the second step, where application uses original request and/or their selected synonyms for advanced search using Google search engine with the aim to extract only that paragraphs, which contain key words from the search-request and to store them in database. Such details as content type (html, pdf, audio, video) will be processed properly via “Content-Type” header of the web-page and web page markup. Thus, we obtain the behaviour model of a user performing manual scraping. n the course of our research we found out that today’s e-learning environment suffers from the cold start problem – lack of ready-made content for courses. At Phase 1 we used synonymous connections to search for dictionaries that are similar in meaning with the help of a crawler, and uses them at Phases 2–4 for advanced search using the Google engine. Our application succeeded in preprocessing searched data and exposed it in html format for user arrangement. The next 3 phases should be designed to present logically connected well merged content and export it in Moodle via Page and File standard plug-ins.