# TongueType > TongueType is a macOS menu bar voice dictation app powered by local Whisper AI. It runs 100% on-device using Apple Silicon, with no cloud processing, no accounts, and no subscriptions. ## Key Facts - Category: macOS voice dictation / speech-to-text application - Price: Free with optional one-time Pro upgrade (no subscription) - Platform: macOS 14+ on Apple Silicon (M1/M2/M3/M4) - Privacy: All audio processing happens locally. No data is sent to any server. - AI Model: OpenAI Whisper via CoreML on Apple Neural Engine - Languages: English, Spanish, French, German, Portuguese, Italian, Japanese, Chinese, Korean, Dutch, Russian (plus auto-detection) - Developer: A Beautiful Site, LLC - Website: https://tonguetype.app ## How It Works 1. Hold your hotkey (Right Option key by default, configurable) 2. Speak in any of 12 supported languages 3. Release the key, and text appears at your cursor ## Features - Real-time voice dictation via hold-to-record hotkey - Audio/video file transcription (WAV, MP3, MP4, MOV) - 12 languages with auto-detection - Smart post-processing with custom rules - Menu bar app with no dock icon and no windows to manage - Works offline after initial model download ## Free vs Pro - Free: All features included, with 30 minutes of live dictation per calendar month and a 10-second cap per file on file transcription - Pro: One-time purchase that removes the monthly limit and the per-file cap. Unlimited dictation and full-length file transcription, no subscription. Check the app for the latest pricing - Refunds: Full refund within 30 days of purchase, no questions asked. Beyond 30 days, refunds are reviewed case-by-case ## Comparison with Alternatives - vs macOS built-in dictation: TongueType is entirely local (Apple's can route audio to servers), has a configurable hotkey, supports file transcription, and has custom post-processing rules - vs cloud-based dictation (Google Docs voice typing, Otter.ai, etc.): No internet required after setup, no account needed, no subscription fees, complete audio privacy - vs other Whisper Mac apps (MacWhisper, Superwhisper): Menu bar simplicity, hold-to-record workflow, free tier available, file transcription included ## Frequently Asked Questions ### Is TongueType really free? Yes. TongueType is free with every feature included. The free tier covers 30 minutes of live dictation per calendar month, and file transcription is capped at the first 10 seconds of each file. TongueType Pro is a one-time purchase that removes both limits for unlimited dictation and full-length file transcription. No subscriptions. No recurring charges. ### Does my audio get sent anywhere? No. TongueType uses Whisper AI running locally on your Mac's Apple Silicon. Your audio is processed entirely on-device. There are no servers, no analytics, no telemetry. TongueType works completely offline after the initial Whisper model download. ### What languages does it support? English, Spanish, French, German, Portuguese, Italian, Japanese, Chinese, Korean, Dutch, and Russian. Plus auto-detection. ### Why does it require Apple Silicon? TongueType runs the Whisper AI model on Apple's Neural Engine via CoreML. This requires an M1 chip or later. Intel Macs are not supported. ### How is this different from macOS built-in dictation? Apple's dictation can send audio to Apple servers and requires enabling in System Settings. TongueType is entirely local, uses a configurable hold-to-record hotkey, supports audio/video file transcription, has smart post-processing with custom rules, and sends zero data to any cloud. ### How accurate are the transcriptions? TongueType uses OpenAI's Whisper model, which handles accents, background noise, and mumbling well. No speech-to-text is perfect, but it is accurate enough that fixing transcriptions takes less time than typing from scratch. ### Can I use it for coding? Yes. It works anywhere your cursor can type: code comments, commit messages, documentation, chat messages. ## Privacy Summary TongueType collects nothing. No analytics, telemetry, crash reporters, or tracking. Audio is captured locally, transcribed by Whisper on-device, inserted at the cursor, and discarded. The only network connections are: downloading the Whisper model from Hugging Face on first setup, and (if you choose to buy Pro) connecting to our payment provider to complete the purchase. ## User Reviews - "My whole job is writing prompts all day. Now I just say them. Faster than typing, and my wrists thank me." — Jamie K., Prompt Engineer - "Hold key. Talk to computer. Words appear. We're living in the future and it costs zero dollars." — Andrew R., Startup Founder - "Finally, dictation that doesn't require an internet connection, a user account, a blood sample, and my firstborn." — Morgan T., Privacy Enthusiast - "I transcribed a 2-hour meeting recording in minutes." — Sam L., Project Manager - "No cloud, no subscription, no account. I keep waiting for the catch. It's been six months." — Larry F., Privacy Advocate - "Finally, a dictation app that doesn't autocorrect 'Kubernetes' to 'cucumbers.'" — Quinn L., DevOps Engineer ## Links - [Homepage](https://tonguetype.app) - [Documentation](https://tonguetype.app/docs) - [Getting Started](https://tonguetype.app/docs/getting-started) - [Dictation](https://tonguetype.app/docs/dictation) - [Transcribing Files](https://tonguetype.app/docs/file-transcription) - [Postprocessing](https://tonguetype.app/docs/postprocessing) - [Settings & Appearance](https://tonguetype.app/docs/settings) - [Free vs Pro](https://tonguetype.app/docs/pro) - [Privacy Policy](https://tonguetype.app/privacy) - [Terms of Use](https://tonguetype.app/terms)