A digital media search and presentation service will receive audio streams from multiple audio content sources and capture text segments, each of which corresponds to words in a snippet of the audio stream. When a client device receives a search request, the service will identify text segments that correspond to the search request. The service will select one or more of the text segments as candidate search results. The service will cause the client device to output user-selectable fields, each of which includes information corresponding to the selected text segments in a user-selectable interface by which a user may select the audio content sources for each of the text segments. When a user selects one of the fields, a media player of the client device will access and play audio content that the audio content source is transmitting at the time.