Added Polish, Czech, Hungarian, Romanian, Ukrainian, Bulgarian, and Croatian localizations, bringing Notelyra to 17 supported languages. This release also refreshes localized App Store metadata and screenshots.

Notelyra 1.1 adds transcript learning memory. This update helps Notelyra improve transcript quality over time based on your own edits and review actions. When you correct a transcript segment, Notelyra can remember that preference locally and apply similar corrections to future lesson recordings after transcription is complete. What’s new: - Transcript learning memory for user-edited corrections - Automatic transcript refinement after transcription finishes - Per-segment transcript editing - A review flow for learned corrections applied by the app - Options to reject, pause, or resume individual learned corrections - A learning memory dashboard showing learned preferences and evidence counts - Recognition language display and quick language switching - Improved handling for music lesson vocabulary and mixed-language music terms - Better playback audio session behavior for louder, clearer recording playback - General stability and transcription workflow improvements Learning memory is designed to stay lightweight and user-controlled. It does not retrain the underlying speech recognition model. Instead, Notelyra learns local correction preferences and applies them as a refinement layer after the system transcript is created.

Notelyra 1.1 adds transcript learning memory. This update helps Notelyra improve transcript quality over time based on your own edits and review actions. When you correct a transcript segment, Notelyra can remember that preference locally and apply similar corrections to future lesson recordings after transcription is complete. What’s new: - Transcript learning memory for user-edited corrections - Automatic transcript refinement after transcription finishes - Per-segment transcript editing - A review flow for learned corrections applied by the app - Options to reject, pause, or resume individual learned corrections - A learning memory dashboard showing learned preferences and evidence counts - Recognition language display and quick language switching - Improved handling for music lesson vocabulary and mixed-language music terms - Better playback audio session behavior for louder, clearer recording playback - General stability and transcription workflow improvements Learning memory is designed to stay lightweight and user-controlled. It does not retrain the underlying speech recognition model. Instead, Notelyra learns local correction preferences and applies them as a refinement layer after the system transcript is created.