At the exhibition, NineNineSix showcased a speech synthesis model called KaniTTS. The developers claim that their technology allows for real-time speech generation with performance three times higher and costs up to ten times lower than similar solutions from global giants like ElevenLabs, OpenAI, and Google. The model is available under the Apache 2.0 license, making it free to use.
KaniTTS boasts impressive technical specifications: it can generate 15 seconds of speech in just one second on a standard NVIDIA RTX 5080 graphics card, simplifying the implementation of the technology without the need for expensive cloud solutions. On the Hugging Face platform, the model has been downloaded over 15,000 times and currently supports eight languages, including Kyrgyz, English, German, and Chinese.
Additionally, the startup presented an automatic speech recognition model called Kyrgyz Whisper, which has been fine-tuned using data from OpenAI. The use of 2,000 hours of recorded Kyrgyz speech has significantly reduced the recognition error rate from nearly 100% to 0.2%, addressing the issue of insufficient support for rare languages in the global market.
Participation in CES was organized by the High Technology Park of the Kyrgyz Republic. According to the PVT, the IT sector of Kyrgyzstan is demonstrating stable growth: over the past five years, the volume of service exports has increased 45 times. In 2024, specialists from Kyrgyzstan earned $130 million in international markets, of which 40% (over $50 million) came from the USA.
