@tokhumcom
Bishkek team open-sources voice AI model that knows when to stay silent
The same team offers AKYLAI, a text-to-speech model for Kyrgyz that people can try, according to a post on X.
Tokhum AI · 1 min read · reviewed by Leyla
Key facts
- A team in Bishkek open-sourced a voice AI model that knows when to stay silent.
- The same post points to AKYLAI, a text-to-speech model for Kyrgyz that readers can try.
- The team and the open-source model were not named in the 21 October 2026 post.
A team in Bishkek has open-sourced a voice AI model that recognises when to stay silent, according to a post on X published on 21 October 2026. The post did not name the team or the model.
The post also points to AKYLAI, a text-to-speech model for Kyrgyz, and links to a page where readers can try it. AKYLAI is described in the post as a bonus item attached to the main announcement.
The post gives no details on the size of either model, the training data behind them, the licence they are released under, or where the open-source model can be downloaded. It says further information is in a carousel of images.
No launch date for either model, no pricing and no usage figures were included in the post.
Related stories
EventProductKazakhstan
Alatau Hub to hold Digital Day for Almaty region IT projects on 23 October 2026
FounderProductUzbekistan
Uzbekistan's Mazza began as a Telegram bot to sell unsold cafe food
ProductKazakhstan
REFACTOR.KZ builds taxi fleet debt system in 10 days
ProductUzbekistan
iDeal launches rental platform for Tashkent tenants and landlords
Central Asia Startup Digest
The region's startup week in one email
Every Thursday: funding rounds, launches and open calls from six countries. In English, free, unsubscribe in one click.
See what's inside →