An audio social networking environment is described that provides a platform for users to generate audio-only content for consumption by one or more other users that may or may not have a social networking relationship with the user creating the audio-only content. Users are able to verbally generate stories using an audio based virtual assistant that receives the stories. The stories are analyzed to identify a tone of the story and one or more categories to associate with the story. The analysis of the story can also include suggesting audio effects to the user for including in the story. When a user requests stories, the user preference information of the requester can be used to identify stories for playback that meet the requesting user's preferences.