Aim of the master thesis is to offer a versatile and accessible system for semi-automatic bibliography generation using algorithmic approach. Bibliography is one of the oldest activities of library science that is encountering new challenges in today's digital age as a result of a multitude of available bibliographic information in various structured bibliographic records and database web interfaces. Bibliographic data can be represented in many formats, which in turn creates problems during the process of creating a bibliography, making it difficult to approach bibliography creation in a fast and accessible way. In addition, structured digital data are meant primarily for machine reading which presupposes algorithmic handling, thus producing another motive for the master thesis. The proposed system for semi-automatic bibliography generation includes features for: reading various formats of bibliographic metadata, including specialized (MARC, BibTeX, RIS, etc.) and general (XML, JSON, etc.) formats; filtering of records according to user-specified parameters (e.g. date of publication, publisher, author, etc.); grouping of records according to user-specified parameters; defining a citation style and exporting the bibliography in formats such as HTML, TXT, PDF or DOCX. The thesis also discusses relevant bibliography topics and reviews the importance as well as advantages and disadvantages of algorithmic approach to bibliography generation. The master thesis represents an improvement in contemporary activity on bibliography with an emphasis on use and adaptation of new technologies as well as a possibility of creating versatile software tools tailored to local needs and tradition.