Learning abstract concepts in data structures and algorithms (DSA) courses is often difficult for students. To improve understanding of DSA topics, numerous algorithm visualization (AV) systems and stand-alone AVs have been developed, supporting a wide range of algorithms and different engagement strategies. Prior studies show that active engagement of students is necessary to make AVs educationally effective. In this paper, we introduce JSAV, a new JavaScript framework for creating engaging algorithm visualizations with active learning features. JSAV is meant to be used with HTML5-based online learning materials. We describe the special features of JSAV that support active learning, and discuss its potential for use in online education projects