Skip to main content

TTS overview

Lumia has a built-in text-to-speech engine that all parts of the system share: alerts, chatbot commands, custom JavaScript actions, the REST API, WebSockets, and plugins. Anything that can produce a string can also speak it.

Sections​

  • Triggering TTS - every surface that can fire TTS: REST API, chatbot, alerts, Custom JS, plugins.
  • Voices - system voices, Lumia cloud voices, marketplace voices, ElevenLabs/TTS Monster plugins.
  • Expressive markup - pause tags, sound tags, paragraph breaks, and punctuation guide.
  • Queue and playback - how the TTS queue works, queue lanes, skip, mute, and waitForTts.

How TTS works​

Every TTS request goes through the same pipeline regardless of where it originates:

  1. The message is resolved through the variable system; {{username}}, {{amount}}, {{counter=deaths}}, etc. are all substituted before speech synthesis.
  2. The voice is resolved: if the request specifies a voice, that wins; if not, the command's configured default is used; if there's no command-level default, the global TTS voice setting applies.
  3. The resolved message is handed to the TTS engine based on the voice source:
    • System: the OS speech synthesizer (no network call, always available).
    • Cloud: Lumia's hosted Polly-backed voices (Brian, Amy, Joanna, etc.).
    • Marketplace: community voices from the Lumia voice marketplace.
    • Plugin: a plugin such as ElevenLabs TTS or TTS Monster intercepts and generates audio.
  4. The audio is dispatched to the configured output device and placed into the TTS queue (or plays immediately if skipQueue is set).

Where TTS appears in the UI​

Every chatbot command, alert, and automation action has a TTS tab. The fields are:

FieldWhat it does
MessageTemplate string spoken aloud. Supports full variable syntax.
VoiceWhich voice to use. Leave blank to inherit the global default.
Volume0–100. Windows only; macOS uses the system volume.
DelayMilliseconds to wait before speaking.
Wait for TTSPause the action pipeline until TTS finishes playing.

Quick paths​

  • Just want to fire TTS from code or an external app? → Triggering TTS
  • Want to pick or add a voice? → Voices
  • Want natural-sounding delivery with pauses and sounds? → Expressive markup
  • Need to control queue order or skip? → Queue and playback
  • Mute all TTS globally? → {{set_sfx_muted=true}} in any template, action, or chatbot command