The fast advance of the technology has originated the proliferation of digital sources of information. This computer evolution has caused the creation of digital libraries that have become a big pillar for the diffusion of knowledge. However, the information contained in the digital libraries is not totally described and its exploitation is still insufficient. Recently, it has been proven that describing the information by using “metadata” can be fundamental for the improvement of the research of the information within a digital library. Our approach is based on the creation and the introduction of new “metadata” able to describe, in our case, the PhD theses of the digital library. These “metadata” correspond to the most important concepts of each one of the theses contained in the digital library. At the moment, manual identification of concepts is a long process that is carried out by a specialist of the area. Therefore, we considered the use of tools to be able to automatically extract concepts. In this article we analyze four tools of NLP(Natural Language Processing) able to automatically extract the key concepts of a corpus. These tools are: (1) TerminologyExtractor of Chamblon Systems Inc., (2) Xerox Terminology Suite of Xerox, (3) Nomino of Nomino Technologies and (4) Copernic Summarizer of NRC. This paper also presents a prototype developed to automatically insert concepts into digital theses.