AI Recorder: Live Transcriber

After many late nights, it’s finally ready. This update brings a more complete on-device recording and transcription experience: - Added Parakeet V3 Near-Realtime captions with on-device support for 25 European languages - Picture in Picture captions now support adjustable window proportions and text sizes, with bilingual previews - Dynamic Island transcripts now offer small, medium, and large text sizes - Nemotron can now detect the spoken language automatically - MOSS on-device speaker refinement now supports both Apple Speech and Nemotron - Added offline speaker recognition to separate speakers and identify people you’ve previously saved - Improved conversation playback by combining adjacent segments from the same speaker into continuous passages - The full transcript editor now links the cursor to the matching audio, so you can listen and edit without losing your place - Added timestamped recording chapters to help you find key moments faster - Added more font and app icon theme options This update also fixes several issues and continues to improve stability and the overall experience. Thank you to everyone for your continued support and patience. LiveTranscriber is still evolving, and some features remain in Beta. There may also be issues we haven’t discovered yet. If you experience problems with recognition, model downloads, recordings, or captions, please let me know through Settings > Feedback. If possible, include the steps to reproduce the issue and relevant screenshots to help me investigate it more quickly. I read every piece of feedback. Thank you for helping make LiveTranscriber better!After many late nights, it’s finally polished and ready. This update brings a more complete on-device recording and transcription experience: - Added Parakeet V3 Near-Realtime captions with on-device support for 25 European languages - Picture in Picture captions now support adjustable window proportions and text sizes, with bilingual previews - Dynamic Island transcripts now offer small, medium, and large text sizes - Nemotron can now detect the spoken language automatically - MOSS on-device speaker refinement now supports both Apple Speech and Nemotron - Added offline speaker recognition to separate speakers and identify people you’ve previously saved - Improved conversation playback by combining adjacent segments from the same speaker into continuous passages - The full transcript editor now links the cursor to the matching audio, so you can listen and edit without losing your place - Added timestamped recording chapters to help you find key moments faster - Added more font and app icon theme options This update also fixes several issues and continues to improve stability and the overall experience. Thank you to everyone for your continued support and patience. LiveTranscriber is still evolving, and some features remain in Beta. There may also be issues we haven’t discovered yet. If you experience problems with recognition, model downloads, recordings, or captions, please let me know through Settings > Feedback. If possible, include the steps to reproduce the issue and relevant screenshots to help me investigate it more quickly. I read every piece of feedback. Thank you for helping make LiveTranscriber better!