Designing artificial intelligence tools dedicated to biosignal data (e.g. electroencephalography, EEG) analysis is an increasingly considered topic, motivated by clinical needs. In this context, data annotation is a crucial task, as a prerequisite to supervised machine learning / deep learning approaches. However, annotation is a tedious, time consuming and error-prone task, that has generally to be carried out in a limited time by clinicians. Based on these considerations, we propose a software tool dedicated to a clinical use, that aims to tackle the main difficulties encountered by clinicians. It is designed as a thin-client, that can be run on a web interface without requiring complex technical dependencies. It allows user-friendly, interactive visualization and annotation, while minimizing the manual interactions. It relies on an extensible data exchange format specifically tailored for storing biosignal data, associated metadata and annotations. A semantic web paradigm is considered for metadata modeling, which allows to aggregate independent features from different sources, users, and to valorize information from different experimental protocols. This strategy allows many users to collaborate on the annotation task whereas reducing their coordination effort and optimizing the quality of the annotations.