– Real-time AI interpretation powered by cutting-edge generative AI models for Japanese–Korean and Japanese–English, available starting today –

Kotoba Technologies Japan, Inc. (Head Office: Chiyoda-ku, Tokyo; President: Noriyuki Kojima; hereinafter “the Company”) announces that it has begun service integration with “Dare Demo AI,” a new service launched by SoftBank Corp. on April 17, 2026.

Through this integration, users can easily experience Kotoba’s real-time interpretation app via “Dare Demo AI.” In addition, customers of SoftBank, Y!mobile, and LINEMO will be eligible for a special promotion: a significantly discounted price for three months when subscribing to the paid plan of the interpretation app.

Coinciding with the launch of this service, the Company has also officially released a Japanese–Korean simultaneous interpretation model powered by its proprietary generative AI model, “Koto.”

■What is “Daredemo AI”?

“Daredemo AI” is a carrier-free service provided by SoftBank that allows users to easily experience popular AI technologies through simple operations. Through this platform, more users will be able to try Kotoba’s real-time interpretation app.

■Overview of Kotoba’s Service on the Platform

 Within “Daredemo AI,” Kotoba offers a “simultaneous interpretation experience.” Leveraging its cutting-edge generative AI models, the app delivers real-time speech interpretation with low latency and high accuracy, enabling seamless communication without language barriers.

Additionally, SoftBank, Y!mobile, and LINEMO subscribers can enjoy a special offer of approximately 94% off the monthly subscription fee for three months.

■Newly Released Japanese–Korean Speech-to-Speech Model

With this launch, the Company has officially introduced a Japanese–Korean simultaneous interpretation model powered by its proprietary “Koto” generative AI model.

“Koto” is an end-to-end generative AI model developed in-house that directly translates speech into speech in real time. This enables both low latency and high accuracy.

Between Japanese and Korean—languages with similar structures—the system achieves an average latency of approximately 1–2 seconds, significantly lower than that of Japanese–English translation, while maintaining high translation quality.

By delivering unprecedented levels of quality in AI-based simultaneous interpretation for Japanese–Korean communication—where effective tools have historically been limited—the Company aims to remove language barriers.

■Comment from the CEO

“We are honored to participate in SoftBank’s ‘Daredemo AI.’ This is a great opportunity to bring the speech AI technology we have been developing for years closer to a wider audience.

We are also confident that the newly announced Japanese–Korean speech-to-speech model represents a new stage in real-time interpretation. We will continue to deliver experiences that eliminate language barriers to as many people as possible.”

— Noriyuki Kojima, CEO, Kotoba Technologies Japan, K.K.

■Future Outlook

The Company continues to develop highly accurate speech AI models centered on Japanese, English, Chinese, Korean, and Spanish.

Within this fiscal year, it aims to further improve the accuracy of its simultaneous interpretation AI to significantly reduce language barriers in business settings.

Starting with its participation in “Dare Demo AI” and the release of the Japanese–Korean model, the Company will continue to drive innovation while addressing user needs across the East Asian language market.

https://prtimes.jp/main/html/rd/p/000000004.000151406.html