On February 19, 2026, Kyrgyzstan’s High-Tech Park announced the success of the NineNineSix startup, which unveiled Kani TTS 2, an advanced speech synthesis model. This innovative technology, now one of the top three text-to-speech (TTS) models on Hugging Face, significantly enhances its predecessor by enabling up to 40 seconds of stable speech generation in a single run. Kani TTS 2 supports multiple languages, including English, Spanish, and Kyrgyz, and features zero-shot voice cloning capabilities.
The NineNineSix team focuses on creating AI solutions for underrepresented languages, emphasizing open-source contributions by providing the complete pre-training code. This allows global researchers to develop tailored voice models. Co-founder Nursultan Bakashov stated that their mission is to prove Kyrgyzstan can produce world-class technology. With approximately 400 million parameters, Kani TTS 2’s architecture is versatile and resource-efficient, making it suitable for various applications in the expanding AI market.
Source link
