Skip to content

Reference

CommandShortcutDescription
Voice Soundboard: Speak Text...Enter text to speak
Voice Soundboard: Speak SelectionCtrl+Shift+SSpeak selected editor text
Voice Soundboard: Stop SpeakingStop current playback
Voice Soundboard: Change VoiceQuick-pick voice selector
Voice Soundboard: Toggle Soundboard PanelShow/hide sidebar
Voice Soundboard: Export Dialogue as WebVTTExport dialogue script as .vtt subtitle file

48 voices across 8 languages:

LanguageVoices
English (US)12 voices (6 male, 6 female)
English (UK)4 voices
Japanese6 voices
Mandarin6 voices
Spanish4 voices
French4 voices
Hindi4 voices
Italian4 voices
Portuguese4 voices

Each voice has a unique ID (e.g., bm_george, af_sarah) used in settings and dialogue scripts.

The Voices tab in the sidebar provides search and filtering:

  • Search — type part of a voice name or ID to filter the list
  • Filter chips — quick toggles for American, British, Japanese, Male, and Female
  • Preview — each voice card has a Preview button that plays a short sample

You can also change the active voice from the Command Palette with Voice Soundboard: Change Voice, which opens a quick-pick list showing every voice with its accent, gender, and style.

The Dialogue tab lets you write multi-speaker scripts:

  1. Write lines in Speaker: Text format (e.g., Alice: Hello there!)
  2. Assign a voice to each speaker using the cast panel that appears below the script
  3. Click Play All to hear the full dialogue sequentially
  4. Click Export VTT to synthesize all lines and save as a WebVTT subtitle file

The exported .vtt file includes:

  • Cumulative timecodes calculated from audio durations
  • Voice tags (<v Speaker>) for each line
  • Standard WebVTT format compatible with video players, subtitle editors, and accessibility tools
SettingDefaultDescription
voiceSoundboard.defaultVoicebm_georgeDefault voice ID
voiceSoundboard.speed1.0Speech speed (0.5 to 2.0)
voiceSoundboard.backendpythonTTS backend (python, http, or mock)
voiceSoundboard.autoStarttrueAuto-start server on activation
voiceSoundboard.formatwavAudio output format (wav, mp3, or ogg)
voiceSoundboard.openaiApiKey""OpenAI API key for cloud TTS (http backend)
voiceSoundboard.pythonPath""Path to Python executable (auto-detected if empty)
BackendHow it worksNetwork
python (default)Kokoro TTS via Python subprocessNo network — fully local
httpConnects to user-configured TTS API (OpenAI, ElevenLabs)Requires API key, sends text to cloud
mockReturns silence — for testingNo network
  • Local backend (default): Kokoro TTS runs entirely on your machine — no network egress
  • Cloud backend (optional): sends text to user-configured TTS API — requires explicit API key setup
  • Audio playback: uses spawn with args array, not shell execution
  • No telemetry is collected or sent