In some implementations, a computing system can reduce startup delays for presenting remote media items. For example, when a user requests media content from a media service, the computing system can include information about the first media item to be played in the initial startup messages to reduce the number of roundtrips needed between client and server devices to present the first media item. A media item can include an introductory asset. The introductory asset can be a portion of the full media item that can be decrypted with a local key or not encrypted at all so that a media client can receive and start presenting the introductory media asset quickly while the full media item is being processed by the media client. In some implementations, the system can remove leading silence from the media item so that the amount of delay perceived by the user is reduced.