Skip to content

Dictate a draft

View Markdown

Turn speech into editable text with an explicitly chosen transcription engine.

Sections in this article

System microphone

Orylo uses the macOS system default microphone for dictation and Narration. To change the input, open System Settings → Sound → Input. Orylo does not provide a separate microphone selector.

Choose the engine and language

Speech language is independent of the interface language. There is no automatic fallback between engines: if the selected engine fails, resolve the error or explicitly choose another.

EngineBehavior
Apple SpeechUses supported speech languages reported by your Mac. Optional live transcription can revise early text while you speak.
Whisper LocalUses the multilingual model you explicitly download and choose. Transcribes after recording stops and supports automatic language detection.

Record and review

  1. Start recording

    Click the microphone button in the Composer you are using. macOS requests microphone permission when needed.

  2. Stop or cancel

    Stop to transcribe, or cancel to discard the recording. While voice input is active, Escape cancels it before closing the window.

  3. Edit the draft

    Review names, numbers, and punctuation in the editable text before sending. Transcription is not automatic submission.

Keep the target clear

Notch Composer and the conversation window have independent recording states. Switching away from a conversation cancels its active recording or transcription instead of writing into another draft.

Whisper preserves the spoken language rather than translating it. A language choice does not guarantee a particular Chinese writing system.

Start here

Use ↑ ↓ to choose, Enter to open, and Esc to close.