Skip to main content
Hold the shortcut, talk, let go. Your words appear wherever the cursor is: an email, an editor, a chat window, a search bar.

​
The dictation loop

1

Put your cursor where you want the text

Any text field in any application.
2

Hold the Dictate shortcut and speak

3

Release the keys

SpeakoFlow transcribes and inserts the text into the active app.
Voice-activity detection filters silence and background noise, so pausing to think does not leave gaps.
Most people speak around 150 words a minute and type around 45. Dictation is roughly three times faster.

​
Hold or tap

The default is Hold: the keys record while held, and releasing stops. Switch to Tap and one press starts recording, the next press stops it. See Hold or tap for details. The toggle is on the Home page (in the shortcuts card) and in Settings → Shortcuts. It applies to every recording shortcut: dictation, dictation with cleanup, ask, and conversation.

​
Dictating into SpeakoFlow’s own window

When SpeakoFlow’s main window is focused and the recording did not start in another app, the text is inserted directly into the field at the caret instead of pasting through the clipboard. This avoids the timing problem where a clipboard paste into the app’s own webview would arrive after the clipboard was already restored.

​
First press before the model is ready

On a new install the speech model downloads while the onboarding tour plays. If you press the dictation shortcut before it finishes:
  • Model still downloading. The overlay shows a pill with the download progress (“Speech model downloading” with a percentage). Dictation starts working when the download completes.
  • No speech model installed. The overlay shows “No speech model yet. Pick one on the Models page.”
If you are using a cloud transcription provider, no local model is needed and the shortcut works immediately. See Cloud transcription.

​
Good to know

The overlay shows “Getting mic ready…” until the microphone starts delivering audio, then switches to a live waveform. That cue comes from real audio frames, not a fixed delay. Once the wave moves, your words are being captured.
A single recording stops itself at 30 minutes and transcribes automatically. Audio is only written to disk after a recording stops, so the cap means a forgotten session still produces a transcript instead of losing everything.
Turn on audio feedback in Settings → Audio → Sounds, then pick a theme: Default, Marimba, Pop, or Click. A Custom option appears if you have supplied your own sound files. Sounds are off by default.“Mute while recording” silences system audio for the length of a recording. Also off by default.
Pick your input in Settings → Audio. Two related options:Always-on microphone (Settings → Audio, off by default) keeps the mic warm so the first words are never clipped, at the cost of a little battery.Clamshell microphone (Settings → Debug) picks a different input for when a laptop lid is closed.

Hold or tap

Record without holding the keys.

Output and pasting

Control how text is inserted.

Custom words

Spell names and terms your way.
Last modified on August 30, 2026