New Speech Engines
Added Supertonic — a high-quality 10-voice pack supporting a wide range of languages: Arabic, Bulgarian, Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, German, Greek, Hindi, Hungarian, Indonesian, Italian, Japanese, Korean, Latvian, Lithuanian, Polish, Portuguese, Romanian, Russian, Slovak, Slovenian, Spanish, Swedish, Turkish, Ukrainian, and Vietnamese.
Added Pocket TTS — an incredibly lightweight Reference Voice engine. It's currently the fastest and lightest speech engine available in the app. Supports English, French, German, Spanish, and Italian.
Added VoxCPM2 — a new speech engine supporting both Reference Voice and Voice Design, with support for 30 languages: Arabic, Burmese, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Indonesian, Italian, Japanese, Khmer, Korean, Lao, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Tagalog, Thai, Turkish, and Vietnamese.
Chinese additionally supports 9 regional dialects: Cantonese, Sichuanese, Wu, Northeastern Mandarin, Henan, Shaanxi, Shandong, Tianjin, and Hokkien.
Voice & Audio
RVC migrated from Python to audio.cpp — should now run faster while using less memory and disk space.
Added new DSP voice-shaping effects, including Auto-Tune.
Updated the Piper voice list, including the new Italian speaker Serena.
Fixed Qwen3 Reference Voice sometimes generating noise during synthesis.
Fixed Qwen3 cutting off the final word of short messages.
Fixed automatic audio saving causing the TTS textbox to lose focus.
Demo Improvements
The Demo can now use Voice Design. Adding, removing, and editing Voice Designs remains restricted.
The Demo can now use 1 Reference Voice.
Text & History Navigation
Added Up/Down arrow history navigation to the TTS overlay.
In the main TTS textbox:
Up/Down navigates history when the textbox is empty.
Ctrl + Up/Down navigates history while text is present.
Install Manager & Localization
Optimized the Install Manager.
Improved localization across the app.
Fixes & Polish
Fixed the Clip waveform selector.
Various minor UI fixes.
Additionally, Text to Sing has made great progress and is available to try on the Preview branch.