Troubleshooting

Open System Settings > Privacy & Security > Microphone and confirm Tota is toggled on.
If using an external microphone, check that it is connected and selected in Settings > Dictation > Audio Input.
Try selecting "Automatic (Built-in Mic)" as the input device.
Tota shows an on-screen alert whenever the active microphone changes, so check which mic the last recording actually used.

Open System Settings > Privacy & Security > Accessibility and confirm Tota is toggled on.
If you just granted the permission, you may need to restart Tota for it to take effect.
Verify that Auto-paste after transcription is enabled in Settings > Dictation > Behavior.
If you switched to a different app while transcription was finishing, this is intentional: the text is on your clipboard and Tota prompts you to press Cmd+V.

Try a larger model. Go to Settings > Models, select a more accurate model (Small, Medium, or Large), and click Load Model.
Speak clearly and minimize background noise.
Check that the correct microphone is selected in Settings > Dictation > Audio Input.
For names, jargon, and technical terms, set up a glossary — see the Glossaries article.

Confirm the wake word is enabled in Settings > Dictation > Behavior.
Open System Settings > Privacy & Security > Speech Recognition and confirm Tota is toggled on.
If macOS speech recognition breaks mid-session, Tota disables the wake word and shows a banner — toggle it back on to retry.

Confirm it is enabled in Settings > Models > Smart Formatting and the model shows Ready.
Right after launch the model may still be loading — Tota tells you when a dictation couldn't be formatted and pastes the raw transcription instead.
The model download is ~2.3 GB; check disk space and internet connection if it fails, then click Retry.

Ensure you have enough disk space for the model download (up to 1.5 GB for Large models).
Check your internet connection — models download from HuggingFace on first use.
Once downloaded, models work fully offline.

Verify that Smart corrections and Automatically learn corrections are enabled in Settings > Corrections.
Accessibility permission must be granted for Tota to read text field edits.
Corrections in Google Docs and similar canvas editors use a clipboard-based fallback (Cmd+A, Cmd+C) which may interfere with your clipboard temporarily.
Words shorter than 5 characters are excluded from learning.

Confirm Monitor clipboard history is enabled in Settings > General > App.
Text entries over 50,000 characters are skipped.
Consecutive duplicate entries are deduplicated.

Verify that the shortcut is not already claimed by another app.
Try resetting shortcuts to defaults in Settings > Dictation > Keyboard Shortcuts > Reset to Defaults.
For hold-to-talk, ensure the modifier is not set to Disabled.

Some apps turn on macOS secure keyboard entry while a password field is focused, and Terminal has an explicit toggle for it in its app menu. While it's active, macOS can pause all global dictation shortcuts — not just Tota's.
Dictation shortcuts resume as soon as the app releases it: click out of the password field, or turn the setting off in the app holding it.
The "Hey Tota" wake word (Settings) is unaffected — it listens to audio, not the keyboard — so it can still start dictation while the banner is up.

This is expected if Hide from app switcher is enabled in Settings > General > App. Toggle it off to restore Cmd+Tab visibility.
The menu bar icon is always accessible regardless of this setting.

Double-check that you entered the correct license key.
Ensure you have an active internet connection for activation.
If you see an activation-limit error, deactivate the license on your old machine first (Settings > About > Deactivate). Contact support if issues persist.

Try Tota today and enjoy 14 days of free squawking