Help center
Set up, troubleshoot, and get the most out of Tonebox.
The full manual ships inside the app — press ⌘? in Tonebox for searchable help that always matches your version. The highlights below cover what early-access users ask most.
Getting started
Install, grant the macOS permissions Tonebox needs, and run your first recording.
- Move Tonebox.app into Applications and open it — no account needed.
- Grant Microphone access; add Screen Recording if you want system audio (the other side of your calls).
- Hit ⌘⇧R to start your first recording, ⌘⇧M to mark a moment.
Hotkeys & dictation
Set up the system-wide push-to-talk hotkey and tune the dictation behaviour to match how you work.
- ⌘⇧D starts push-to-talk dictation into whatever app is focused.
- Every shortcut is remappable in Settings → Hotkeys.
- Voice modes (translate, email reply, standup…) each carry their own prompt.
Transcription & summaries
Pick a Whisper model size, enable CoreML acceleration, and bring your own LLM for summaries and asks.
- Settings → Transcription picks the Whisper model — large-v3 for accuracy on Apple Silicon, base.en for speed.
- Settings → AI wires up Anthropic, OpenAI, or a local Ollama endpoint.
- Summary, task-extraction, and Ask prompts are all editable.
Privacy & data
How Tonebox stores data on your Mac, what (if anything) leaves your machine, and how to back up or delete sessions.
- Sessions live in ~/Documents/Tonebox — plain files you can back up or delete.
- Nothing is uploaded unless you explicitly pick a cloud LLM.
- Export any session to ZIP (audio + transcript + notes) from the session menu.
Still stuck?
Send a note via the contact form — during early access, a human reads every message.