- Live Transcriptions! - Record and transcribe in real time. Pause and resume recording sessions across multiple takes, all in one waveform timeline. - 5 New Local Models - Cohere Transcribe (14 languages), Parakeet Japanese, Parakeet v3 INT4, SenseVoice, and an 83-language Nemotron Multilingual streaming model. - Rebuilt Model Download System - Model downloads are more reliable, with fixed cache path bugs, automatic cleanup of stale files, and better handling of interrupted downloads. - Performance & Memory - Background transcript search, Cohere now routes directly to the Apple Neural Engine, and long audio files process one chunk at a time instead of all at once. - Diarization Quality Improvements - Fixed speaker segment filtering bugs that caused spurious short segments in speaker timelines - Many Bug Fixes

Bulk transcription is here: add multiple files at once and Whisper Snapper will transcribe them in a row, with per-file progress tracking and the ability to add more files even while transcription is running. Two new local models are available: Parakeet TDT-CTC-110M, a faster and more accurate English model (Pro), and Qwen3 ASR, a multilingual model supporting 30 languages including Arabic, Cantonese, and more. Speaker identification now works with all local Whisper models, not just Parakeet. The upgraded diarizer also increases the maximum number of identifiable speakers from 4 to 10. Language Selection: Manually select your language to override auto-detection for accurate transcriptions in your language.

Synced video player: - watch video side-by-side with your transcript Chapters: - auto-generate chapter markers with on-device AI or manually generate, exported in all formats Cleaner UI: - New icon tabbed sidebar for quick access to AI tools, chapters, and options Performance Improvements: - Background model preloading for instant transcription start, AI Translation significantly more reliable Onboarding - Improved intuitive onboarding experience for new users Accessibility improvements: - Dynamic Type and reduced motion support

Enhanced Speaker Separation - VAD-enhanced processing for better speaker separation accuracy Noisy Environment Filter - Improved accuracy by filtering background noise during transcription Timeline Editor Improvements - Resize segments from any edge, smooth auto-scroll during playback, undo/redo actions, and pinch-to-zoom on trackpad Manual Speaker Management - Manually add new speakers to your transcriptions Performance Optimizations - Faster loading, reduced memory usage, and concurrency improvements

Waveform Timeline Editor Visually edit segment timing with draggable boundaries on the audio waveform. Word-Level Playback Words highlight in real-time as they're spoken during playback. Inline Transcript Editing Click any text to edit directly—no mode switching required. UI/UX Improvements Enhanced speaker names, fonts, spacing, and colors with dropdown menus for easy editing. Improved Timing Accuracy More precise playback sync across all models.

Bug Fixes: - Fixed audio player continuing to play previous transcript when switching between transcripts - Fixed compatibility issue with latest MLX framework Improvements: - Added model name display to transcript details (shows which transcription model was used) - Improved accessibility with VoiceOver labels and keyboard shortcuts (⌘O to open, ⌘, for settings) - Added Reduce Motion support for users with motion sensitivity - Added Support link in Settings

Ask and you shall receive! What's New in Version 1.1: • Local AI Summaries — Summarize transcripts using on-device AI • Local AI Translation — Translate to 20+ languages, fully offline • Llama 3.2 Integration — Privacy-first AI that runs entirely on your Mac. • Faster Search — Improved performance for long transcripts • Better Accessibility — Enhanced VoiceOver support