Tap, speak, done. Tonota turns your voice into text entirely on your iPhone — organize it into folders, search every word, and keep it all off the cloud.
Scan to openFree · iPhone 12 or later · iOS 17+
Features
Translate a transcript or run built-in AI tools (summarize, extract key points, and more) — all locally. Pick an MLX model or, on supported devices, Apple Foundation Models with no download. Results stream in, the model name is shown, and reasoning stays collapsed by default. Build up to 5 of your own custom tools.
Record right from your wrist — no need to reach for your phone. Works offline; recordings sync to iPhone and transcribe automatically. Browse recent memos and play audio (play/pause, ±15 s) on the watch.
All transcription runs locally using WhisperKit and Apple's Neural Engine. Your audio and text never touch a server — no account, no tracking, works offline.
Long-press the app icon for a quick-action shortcut, or say "Hey Siri, start recording with Tonota" — no unlocking, no navigating, just speak.
Add a Tonota button to your lock screen or Control Center. One tap starts recording instantly — even before you unlock your phone.
Every recording is kept. Tap play in the detail view to hear the original audio any time — perfect for double-checking a tricky word.
Tap search in any list to filter instantly by title and transcript content — so the memo you need is always a few keystrokes away.
Sort recordings into your own folders and browse by collection. Each memo shows its folder tag in the list so you can see at a glance where everything lives.
Multi-select recordings to move them to any folder or delete them all at once. Keep your library tidy without touching each memo one by one.
Turn on export in Settings and Tonota saves each transcript as a .txt file bundled with the original audio recording to any folder you choose.
English, Chinese, German, French, Spanish, Italian, Japanese, Korean, Portuguese and more. WhisperKit detects your language automatically — or pin one in Settings.
Change the interface language between English, Deutsch, and 中文 without leaving Tonota — no restart needed. Follows your system language by default. Suggested by Nicky Xu.
Pause, resume, or cancel a download mid-way. Friendly error messages with a one-tap "Check network" button when things go wrong. No more phantom progress when you navigate away. Suggested by Kenny Wang.
Tonota was built on one principle: your recordings and transcriptions never leave your phone.
Languages
WhisperKit auto-detects your language. You can also pin one in Settings.
Open Source & Models
Tonota uses open-source on-device AI libraries and downloadable local models. Their notices stay available for review.
Requirements
Tonota is built by one developer. Your reports directly shape what gets fixed and what gets built next.
Free on the App Store. Your first transcription is seconds away.
