An IP telephone which supports a voice conference call with a plurality of other IP telephones over a packet-switched network is disclosed. The IP telephone exchanges audio and signaling messages with the other IP devices during the conference call, and thereby allows three or more conference call participants to talk together over the packet-switched network. The IP telephone includes a memory, which stores a conferencing module and a channel establishment module, and a mixing module for mixing input signals received at the IP telephone to produce a combined signal output. When played back, the combined signal output allows the user to hear what was said by the other conference call participants.