Much recent research deals with the automatic extraction of multimedia metadata. While automatic extraction is already well understood for low-level features (e.g., color), it remains an open issue for high-level (semantic) features [1]. In this context, an evaluation of currently available MPEG-7 annotation tools and frameworks has been performed [2]. Most tools can extract low-level features automatically, and some also provide functionalities to describe highlevel features manually. This paper highlights VAnalyzer1, an MPEG-7 based semantic video annotation tool, which is also able to extract certain semantic annotation automatically. Here, we only introduce the features of the VAnalyzer without discussing its architecture. Finally, the demonstration procedure will be highlighted.