An audio management apparatus manages simultaneous streams of packet-switched audio data for a network conference tool, and allows the combining of conferences that are not co-located. The apparatus uses at least two audio bridges, each of which receives audio inputs from a plurality of local connection points, does mixing of the audio data for each local connection point and provides the audio outputs thereto. Also provided is a communications link between the two audio bridges, and each of the audio bridges creates an audio mix from its local connection points and outputs it over the communications link to the other audio bridge. Each bridge may treat the audio data from the other audio bridge as it does any other call. In this way, the combining of the two conferences is accomplished in a simple and straightforward manner.