Abstract — Social Media are nowadays one of the most used channel for communicating with other people, continuously overwhelmed by a huge quantity of User Generated Content, including text, images and videos. A peculiar aspect of the published material is the variety of its contents, spread by users as if they were Social Sensors. During emergencies people tend to report facts, descriptions, photos, etc. since they’re eyewitnesses of the unfolding crisis. Such information allows machine learning algorithms to automatically detect ongoing emergencies using textual information and to support rescues in preparation for a better response. We developed a tool which offers users a set of functionalities for emergency detection and monitoring using Social Media data captured from Twitter. We describe the modular software architecture of the tool and the techniques adopted to filter messages and use their content for detecting and monitoring unpredictable events. We present and discuss the results of real-world experiments obtained with the tool. While the initial focus of our work was on earthquakes, the tool can be easily reconfigured to support other kinds of emergencies like floods, wildfires, storms, etc. Finally, we introduce some possible functionalities to expand the analysis to multimedia information.