The goal of our research is to meet the growing needs of extraction of relevant information being faced by many disciplines. We present here a brief demonstration of UNITEX/GRAMLAB a software developed at the University of Paris-Est Marne-la-Vallée. It is a multilingual corpus analyzer which is based on a symbolic approach and uses linguistic resources, electronic dictionaries and local grammars. We only focus on some of the main features of UNITEX/GRAMLAB and we do not exhaustively explore their fields of application, especially to develop information extraction systems running over large-scale corpora.