Voice & Audio
Moved Kitten synthesis to the managed KittenSharp ONNX runtime.
Added bundled Resemble Denoise for offline background-noise removal in Sound Editor and Reference Voice.
Updated the Piper catalog with the new Japanese hi-fi Captain voice.
Improved Piper reliability when launched outside the application folder and fixed multi-codepoint phoneme handling.
Settings & Workflow
Install Manager Redesigned - Merged Reference Voice Install and Install Manager dialogs.
Moved Voice Shaping into its own Settings tab.
Added Voice Shaping presets.
Added Reference Voice engine selection in Voice settings.
Added more Appearance presets with live preview.
Unavailable AI Assistance models are now visibly disabled and automatically fall back to an installed model.
Soundboard & Playback
Added 12 default soundboard clips, including applause, rimshot, airhorn, cricket, golf clap, quack, and more.
Simplified Soundboard row actions to a compact icon-only Assign Key button.
Input, UI & Localization
History actions — Added inline Play, Copy, Save, and Remove actions to history entries.
The fullscreen TTS overlay now has a real caret with left/right movement, Delete, Backspace, paste, and Unicode-safe editing.
Improved history navigation and remapped arrow-key handling across normal and fullscreen TTS input.
Application-owned context menus now follow display-scale settings.
Improved hover, focus, selection, pressed, and disabled-state contrast across dark and light themes.
Reduced redundant tooltips while retaining contextual explanations.
Expanded localization coverage for dynamic RVC controls, statuses, settings, and supported languages.
Added low VRAM warning before installing models unfit for local machine.
Fixes & Stability
Fixed clip-recording crashes when a WASAPI audio endpoint is unavailable.
Improved bundled Python worker packaging checks so incomplete installations fail earlier with clearer diagnostics.
Various Settings, layout, lifecycle, cancellation, and localization fixes.