Local-Only Processing
All transcription happens entirely on your Mac. No audio or text is ever sent to any server. Whisper models run on-device using WhisperKit, Smart Formatting runs a local language model via MLX, Audio Drop's speaker diarization runs on-device with Core ML, and the "Hey Tota" wake word uses Apple's on-device speech recognition.
No Telemetry
Tota does not collect, transmit, or report any usage data, analytics, or telemetry. All statistics (word counts, recording duration, etc.) are stored locally and never leave your machine.
Hardened Runtime
Tota is built with macOS hardened runtime enabled, meeting Apple's security requirements for notarized apps.
Permissions
Tota only requests the permissions it needs:
- Microphone — for voice recording.
- Accessibility — for cursor detection, text insertion, and (when Spell what's on your screen is enabled) reading the visible text of the window you're dictating into at the moment you start. That text only biases spelling for the current dictation — it is never stored and never leaves your Mac.
- Speech Recognition (optional) — only requested if you enable the "Hey Tota" wake word.
No additional permissions are requested. Tota does not use app sandbox (required for Accessibility API and global hotkey access), but all network communication is limited to license validation, update checks, and one-time model downloads.
Data Storage
All user data is stored locally in:
~/Library/Application Support/Tota/| File | Contents |
|---|---|
transcriptions.json | Active transcriptions (last 3 days) |
transcription_stats.json | Usage statistics |
dictionary.json | Learned word corrections |
clipboard_history.json | Clipboard history entries |
filler_words.json | Filler word configuration |
snippets.json | Text snippets |
vocabulary.json | Glossaries and their terms |
appGlossaryRules.json | Per-app glossary auto-pick rules |
voiceCommands.json | Voice command definitions |
commandMacros.json | Macro sequences |
deletedBuiltInCommands.json | Built-in commands you've deleted |
license.json | License data |
archives/transcriptions/ | Archived transcription files |
