In this paper, we introduce the construction of our VoiceXML-compliant voice browser that is able to interpret and execute VoiceXML documents. The platform of our voice browser contains a VoiceXML interpreter, a speech recognition (SR) engine, a text-to-speech (TTS) engine, and a computer-telephony-integrated (CTI) user interface. We adopt an open source called OpenVXI as the VoiceXML interpreter. Our main effort for developing the voice browser is to incorporate OpenVXI with our SR, TTS, and CTI modules. In order to test whether our voice browser conforms to the VoiceXML specifications we build three VoiceXML applications including stock quotes querying, automatic call-transfer, and news-reading systems. Besides, we develop a system that allows users to register their own VoiceXML documents in our voice gateway system via Web. They may then call to our voice gateway system and browse their VoiceXML documents by using our voice browser.