Skip to main content

NineNineSix

Category: Voice / Speech AI

A high-performance AI voice technology startup specializing in real-time, expressive text-to-speech (TTS) and automatic speech recognition (ASR) with a focus on underrepresented and localized languages. NineNineSix was founded in 2025. The company is led by Nursultan Bakashov. Based in Bishkek, Kyrgyzstan. Team size: 3-10. Total funding raised: $0. Latest round: Self-funded (Bootstrapped, 2025-01) - NineNineSix. Key investors include ["Nursultan Bakashov (Self-funded)"].

Founded
2025
Headquarters
Bishkek, Kyrgyzstan
Team size
3-10
Total funding
$0

Value proposition

Delivers open-source, enterprise-grade voice synthesis that rivals global tech giants in quality while offering superior computational efficiency and linguistic accuracy for localized markets.

Products and solutions

["KaniTTS (High-fidelity, real-time open-source speech synthesis model)","Kyrgyz Whisper (Advanced automatic speech recognition model)","Arabic-speaking AI Voice Model","Localized AI Assistant Integration (via AkylAI collaboration)"]

Unique value

Successfully bridges the gap between high-end generative AI and underrepresented languages (such as Kyrgyz and Arabic) through an open-source framework that matches the performance of closed-source enterprise solutions.

Target customer

Content creators, software developers, enterprise localization teams, and organizations focused on digital accessibility and language preservation.

Industries served

["Artificial Intelligence","Media & Content Creation","Education & EdTech","Healthcare (Accessibility)","Enterprise Software"]

Technology advantage

Proprietary optimization that reduced Kyrgyz Word Error Rate (WER) from 100% to 0.2%; high-performance inference capabilities that allow for real-time, expressive voice synthesis on standard hardware.

How they differentiate

Focuses on high-fidelity speech synthesis for underrepresented and low-resource languages (e.g., Kyrgyz, Arabic) with an open-source model (KaniTTS) that matches enterprise quality while maintaining high computational efficiency for real-time inference.

Main competitors

["ElevenLabs","OpenAI (Whisper/TTS)","Fish Audio","Cartesia"]

Key partnerships

["Kyrgyz AI Research Institute","AkylAI (First Kyrgyz AI assistant)","Innopolis University (Research collaboration)","Kyrgyz Republic Ministry of Economy (CES Delegation partner)"]

Notable customers

["AkylAI","Kyrgyz AI Research Institute","Independent Content Creators","Local Enterprise Developers"]

Major milestones

["Unveiled KaniTTS at CES 2026 in Las Vegas","Achieved 15,000+ downloads on Hugging Face within the first months of release","Developed the first open-source Arabic-speaking AI voice model built in Kyrgyzstan","Successfully reduced Kyrgyz speech recognition WER to 0.2% using Kyrgyz Whisper"]

Growth metrics

$330K ARR (as of 2025); 15,000+ downloads of KaniTTS on Hugging Face; Kyrgyz Word Error Rate (WER) reduced from 100% to 0.2%.

Market positioning

Specialized AI provider for localized speech technology and open-source high-performance TTS.

Geographic focus

Central Asia, Middle East, and Global (via Hugging Face community)

Patents and IP

No registered patents disclosed; utilizes a strategic open-source IP model to drive global adoption (15,000+ downloads on Hugging Face).

About Nursultan Bakashov

Co-founder of AkylAI (the first Kyrgyz AI assistant), Director of the Kyrgyz AI Research Institute, and a prominent educator in Data Science and Machine Learning in Central Asia. He led the development of the KaniTTS open-source model and has a background as a researcher at Innopolis University.

Official website: