First: amazing project. This is the closest thing I’ve found to a true local AI audiobook reader for macOS.
I tested Russian EPUB/PDF playback.
Current behavior:
Russian text loads correctly
chunking works correctly
playback queue works correctly
audio generation starts successfully
but pronunciation becomes gibberish
This suggests the Kokoro runtime is working, but Cyrillic/Russian phonemization/G2P routing is missing or English-only.
Would it be possible to:
Add Russian/Cyrillic language support
Support external language packs / multilingual Kokoro models
Allow users to drop additional voice/model packs into the kokoro-model folder
Add configurable preprocessing/transliteration hooks before synthesis
The current architecture already feels very close to a perfect local audiobook solution. Russian support would make it incredible for multilingual users.
Thanks for building this.
First: amazing project. This is the closest thing I’ve found to a true local AI audiobook reader for macOS.
I tested Russian EPUB/PDF playback.
Current behavior:
Russian text loads correctly
chunking works correctly
playback queue works correctly
audio generation starts successfully
but pronunciation becomes gibberish
This suggests the Kokoro runtime is working, but Cyrillic/Russian phonemization/G2P routing is missing or English-only.
Would it be possible to:
Add Russian/Cyrillic language support
Support external language packs / multilingual Kokoro models
Allow users to drop additional voice/model packs into the kokoro-model folder
Add configurable preprocessing/transliteration hooks before synthesis
The current architecture already feels very close to a perfect local audiobook solution. Russian support would make it incredible for multilingual users.
Thanks for building this.