Tags: infiniV/VoiceFlow
Tags
fix: Qt threading crash and custom hotkey support - Fix crash when pressing hotkey caused by Qt threading issue - Use Qt Signal/Slot mechanism for cross-thread UI updates - Add custom hotkey capture with modifier-only support (e.g., Ctrl+Win) - Add hotkey validation on backend - Bump version to 1.3.1
fix: model download HuggingFace link and add clear cache option - Fix "Download from HuggingFace" button using api.openExternalUrl() instead of window.open() which doesn't work in Pyloid/QtWebEngine - Add clear_model_cache() method to ModelManager for deleting only faster-whisper models - Add clear_model_cache RPC endpoint in server.py - Add clearModelCache() API method in api.ts - Add "Delete AI Models" checkbox option in Settings danger zone dialog - Only deletes VoiceFlow's faster-whisper models, not other HuggingFace cached models
feat: expand model support with all faster-whisper variants - Add all faster-whisper models: English-only (.en) and distilled variants - Add detailed model metadata (speed/accuracy ratings, sizes, descriptions) - Update turbo model to use official Systran repo - Fix TypeScript type error in ModelDownloadProgress onComplete callback - Enhance Onboarding model selection with category info and best-for guidance - Bump version to 1.2.1