ToSpeak G2, a new text-to-speech (TTS) middleware that can provide users with a wide choice of voices in producing speech contents. ToSpeak G2 features a quick voice production system, and the output voice data of only several megabytes in size can be easily sent via the Internet and installed in an embedded system with a small memory footprint. Using this quick voice production system, a trial version of a new voice can be built within several hours from speech data recorded for several tens of minutes. A new voice can also be created by a combination of two different speakers’ voice components, or by changing the voice parameter settings. ToSpeak G2 makes it possible to offer individuality to various application systems with a wide variety of voices.