Messages, such as voice messages, are received at a server and processed to generate extracted data that includes text from the message, envelope information, or contact information. The extracted data is transmitted to a device that displays a visual representation that includes some of the extracted data. The visual representation (e.g., an icon) enables the user to visually see what is in a voice message. By selecting a particular visual representation, the corresponding message can be retrieved from the message server and played at the device. In some instances, the message server also transmits an initial segment to the device that can be played by the device as well.