Skip to main content

Dictation & voice memos — Regression Test

Prerequisites​

  • Dictation enabled and the Accessibility (Dictation) permission granted
  • A text field in any app focused as the paste target

Steps​

  1. Confirm dictation is enabled in Settings → Voice Dictation, and note the current shortcut.
    • Enable voice dictation is on
    • The Dictation shortcut is shown
  2. Focus a text field, then press-and-hold the dictation shortcut and say a short phrase (e.g. "hello world"), then release.
    • A dictation status window appears with a waveform while held
    • After release it shows "Processing…", then the cleaned-up text is pasted/copied
  3. Press-and-hold again and dictate a longer passage (15+ seconds), then release.
    • The longer text is transcribed and formatted (capitalization/punctuation)
  4. Open Voice Memos from the sidebar.
    • The dictations you just made appear as memos, grouped by day (Today)
    • A memo can be copied and deleted from its ⋯ menu (delete asks to confirm)
  5. Open the Dictionary tab, add a custom word, then delete it.
    • The added word appears with a "Manual" badge
    • Deleting removes it from the list

Variations​

  1. In Settings → Voice Dictation, turn Clean up dictated text off, then dictate the same short phrase.
    • The raw transcript is pasted without LLM cleanup
    • Turn it back on and re-dictate → cleaned-up text returns
  2. Disconnect from the internet and dictate.
    • With the formatter on but backend unreachable, the raw transcript is still pasted (graceful fallback); if transcription itself can't run, a "Dictation failed" error with retry is shown
  3. Change the dictation shortcut: click Record shortcut, press a new valid combo, then use it.
    • The new shortcut is saved and triggers dictation
    • Reset to default restores the original shortcut