AI Project

S'gaw Karen TTS/STT AI Model

Preserving language through technology — building text-to-speech and speech-to-text AI for the S'gaw Karen community.

Project Timeline

1
Formation

January 2026

Karieng Language Service founded with the vision of bridging communication gaps for the Karen community in Australia and beyond.

2
Voice Data Collection App

February 2026

Built an Android application that will allow volunteers to record their voices as they are reading Bible texts

3
Voice Data Collection App Released

February 2026

Released the application on Google play and start recruiting volunteers to start making contribution

4
Collect 1000+ hours of voices

31 March 2026

To be able to train a decent STT model, we will need at least 1000 hours of recordings

5
Train

1 April 2026

Began research and development of the S'gaw Karen TTS (Text-to-Speech) and STT (Speech-to-Text) AI model, partnering with linguists and native speakers for data collection.

6
TTS/STT App Pre-release

Mid April 2026

Release a web site for people to try out generating voices (TTS) and texts (STT).

7
TTS/STT App Release

May 2026

Aim to release the Text-to-Speech and Speech-to-Text web and API service.

8
Start the Translation AI Project

June 2026

Once the TTS and STT are good to go, the next phase is to start building training dataset for English to Karen training.

9
100,000+ Sentences to Start Training

June 2027

Once we have 100,000+ sentences translated it is time to train the AI.

10
First Version API Release

January 2029

Once we are happy with the quality of the translation, we will release the application.