The last decade has seen a rapid development of the number of NLP tools which have been made available to the community. The usability of several e-lexicography tools represents a serious obstacle for researchers with little or no background in computer science. We present in this paper our efforts to overcome this issue in the case of a machine learning system for the automatic segmentation and semantic annotation of digitised dictionaries. Our approach is based on limiting the burdens of managing the tool's setup in different execution environments and lightening the complexity of the training process. We illustrate the possibility to reach this goal through the adaptation of existing functionalities and through using out of the box software deployment technology. We also report on the community's feedback after exposing the new setup to real users of different professional backgrounds.