Many multimedia visualizations abstract the underlying content into aggregate displays, requiring user interaction in order to expose the original text, images or video. The drawback of this approach is that unlike traditional, numerical data, multimedia data is readily interpretable. Users can catch an image or phrase out of the corner of their eye an immediately understand the content. However, this passive discovery cannot take place when content is only exposed through direct interaction. In this paper, we present the Panta Rhei, a peripheral display designed to avoid this pitfall by surfacing original content when the user is not actively engaged with the application. We provide the full implementation details, including the many ways in which the underlying parameters can be tuned to suit various objectives. Since the display can easily support text, images or videos, our goal is to enable more widespread discussion and experimentation involving this technique for multimedia visualization.