TypeWhisper：Mac 本地语音转写

Local speech-to-text for macOS on-device AI, fully private, optional cloud

TypeWhisper for Mac Speech-to-text and AI text processing for macOS. Transcribe audio using on-device AI models or cloud APIs (Groq, OpenAI, xAI/Grok), then transform the result with reusable workflows. Your voice data stays on your Mac with local models - or use cloud APIs for faster processing. TypeWhisper 1.7.0 is the current stable release for macOS. It adds iCloud sync for History and Inbox, a visual keyboard shortcut editor, Undo Last Dictation and Restore Raw Transcript, new indicator styles, expanded dictation recovery, Finder transcription, and compatible local model imports. Development on main targets 1.8; daily builds use the v1.8.0-daily.* line. See the 1.7.0 release notes, release readiness guide, support matrix, and release validation process for the shipped feature set and ongoing 1.x maintenance gates. Manual ship checks are tracked per candidate with the release checklist issue template. typewhisper-demo-full-with-voiceover.mp4 Screenshots The localized macOS screenshot workflow is documented in docs/screenshot-automation.md. What's New in 1.7 iCloud sync and a redesigned History - History and Inbox sync between your Macs through TypeWhisper's private iCloud container, and History groups entries by device with filters and search Undo and recovery - Undo Last Dictation and Restore Raw Transcript are available from the menu bar and as global hotkeys, and Dictation Recovery keeps the last three successful dictations for up to 24 hours so an incomplete provider response can be retried Shortcuts and cancellation - Shortcuts are set on a visual Mac keyboard that follows the active input source and shows conflicts, and cancellation can use double Escape, single Escape, immediate cancellation, or be disabled Indicator themes - Classic, Glass, and Light indicators with a live preview on the new Appearance settings page Import and export - Vocabulary import from Wispr Flow, Handy, and compatible CSV files, export or deletion of all app data in Advanced settings, and settings backups from the CLI with typewhisper export and typewhisper import More ways to transcribe - File transcription from Finder, web media through the optional Web Link plugin, custom local speech models from Hugging Face or a local folder, the Canary ASR plugin, and Parakeet Ultra Text handling - A minimum threshold for number formatting, dictionary corrections before workflow LLM processing, and segmented processing for long dictations Features Transcription Local and cloud engines - Choose from WhisperKit, Parakeet (Ultra, TDT v3 and v2, Fluid Speech Mini, and Nemotron 3.5 ASR Streaming, which shows text while you speak), Apple SpeechAnalyzer, Granite Speech, Qwen3 ASR, Voxtral, Cohere Transcribe, GigaAM (Russian), Groq Whisper, OpenAI Whisper, Soniox, Smallest Pulse, xAI/Grok STT, OpenAI Compatible, and additional bundled or community providers On-device or cloud - Keep audio on your Mac with a local engine, or explicitly configure a cloud API for faster processing and additional model choices Streaming preview - See partial transcription in real-time while speaking (WhisperKit) Short-clip handling - Better retention of brief utterances and fewer false no-speech discards File transcription - Batch-process multiple audio/video files with drag & drop, or start from Finder Custom local models - Import compatible speech models from a Hugging Face repository or a local folder in the settings of local plugins Subtitle export - Export transcriptions as SRT or WebVTT with timestamps Dictation System-wide - Push-to-talk, toggle, or hybrid mode via global hotkey, auto-pastes into any app Modifier-key hotkeys - Use a single modifier key (Command, Shift, Option, Control) as your hotkey Last-transcription actions - Copy or paste your latest transcription with configurable global hotkeys Undo and raw restore - Undo Last Dictation removes the last inserted dictation, and Restore Raw Transcript replaces it with the unprocessed text, without touching the clipboard Dictation Recovery - Keeps failed recordings and the last three successful dictations for up to 24 hours so they can be retried Cancellation - Cancel with double Escape, single Escape, or immediately, or turn cancellation off Visual shortcut editor - Set shortcuts on a Mac keyboard that follows the active input source and shows conflicts and unavailable keys Indicator styles - Choose Notch, Overlay, or Minimal in Classic, Glass, or Light, with optional live transcript preview where supported Sound feedback - Audio cues for recording start, transcription success, and errors Microphone selection - Choose a specific input device with live preview and improved recovery after route changes AI Processing Workflows - Build reusable transformations for translation, rewriting, extraction, formatting, and app-specific automation. Workflows can run automatically by app, website, or app + website combinations, from a dedicated hotkey, as a global fallback, or manually from the Workflow Palette. Hotkey workflows can either start dictation or process the current selection/clipboard directly. LLM provider fallbacks - Order Apple Intelligence (macOS 26+), Groq, OpenAI / ChatGPT, xAI/Grok, Gemini, OpenAI Compatible, and local providers in one global provider/model list. Prompts and workflows inherit that order by default; a workflow with an explicit provider stays on that single provider Speech providers - System voices, xAI/Grok TTS, and experimental local Supertonic TTS can provide spoken feedback and readback Local prompt processing - The Local LLM (MLX) plugin runs on-device on Apple Silicon. It recommends the Gemma 4 E2B/E4B 4-bit models and offers Qwen3.5 2B and LFM2.5 2.6B as experimental lighter models Translation - Translate transcriptions on-device using Apple Translate Personalization Workflow triggers - Per-app, per-website, combined app + website, hotkey, global fallback, and manual palette-only triggers for language, task, engine, prompt, and auto-submit behavior. Web