Selected for the Second Phase of the Domestic Generative AI Development Project "GENIAC" & Beta Release of the Advanced Japanese Speech Generation Model Playground, "Kotoba SpeechGen"

On October 28, we released the beta version of the cutting-edge Japanese speech generation model playground, "Kotoba SpeechGen". The main features of this playground are as follows:

1.     Enhanced Naturalness and Fluency:
Leveraging Preset Voices, it enables the generation of fluent and natural Japanese speech that surpasses traditional text-to-speech (TTS) systems.

2.     Voice Cloning Functionality:
A new feature allowing users to upload their own voice or audio files for cloning.

3.     Multilingual and Real-Time Improvements:
Continuous enhancements to real-time performance, with plans to support English and Southeast Asian languages in the future.

While the playground is not available for commercial use, we are planning to sequentially release APIs for commercial applications. Access to the playground will be granted based on the waitlist order. You can register for the waitlist here.

Additionally, our company has been selected for the second phase of the domestic generative AI development project "GENIAC," promoted by the Ministry of Economy, Trade, and Industry (METI) and the New Energy and Industrial Technology Development Organization (NEDO). This marks our second selection, following the announcement of Phase 1.5 earlier this year in February.
We are committed to advancing technological development by leveraging large-scale computational resources to build a robust foundation for speech generation AI in Japan and Asia.

The selected organizations are as follows：https://www.nedo.go.jp/news/press/AA5_101790.html

Press Release : https://prtimes.jp/main/html/rd/p/000000001.000151406.html