Free-Comment (FC) is a sensory characterization method where panelists are invited to describe the products using their own term by answering open-ended questions. Despite its benefits, notably its spontaneous and natural aspects, the method remains currently underused in practice. The main reason for that is likely to be the cumbersome and time-consuming extraction of relevant sensory information from the text. This presentation introduces a web-based application (accessible from any web browser) that fully automatizes the data analysis of FC data without requiring any coding. The relevant sensory information is extracted from the text using NLP tools combined with semantics. It results in descriptors classified and hierarchized according to an ontology. Then, an algorithm aggregates the descriptors having the same semantic meaning (same parent in the tree structure of the ontology) AND similar profiles of citations that maximize product discrimination. The descriptors are then encoded as check-all-that-apply data and analyzed using multiple-response chi-square related methods.