A method and system for synchronizing in real-time the voice media of a conversation conducted over a network between a first communication device and a second communication. The method includes at each of the first and second communication devices progressively storing in first and second storage elements and transmitting the voice media created using the first and second communication devices to the other communication device respectively. Both the first and second communication devices store in the first and second storage elements the progressively received media from the other device respectively. A mechanism to continually review, ascertain and request the media stored in the first storage element, but not the second storage element, and vice-versa is provided to ensure that the two storage elements contain the same voice media. As a result, the first and second storage elements each maintain real-time synchronized copies of the voice media of the conversation respectively.