Skip to content

Text to Speech

Kristopher Barrantes edited this page Oct 30, 2026 · 5 revisions

Text to speech

Cuebot reads chat, events and anything your actions say aloud. Everything is under Audio > Speech.

Engines

Open Engines to set up the voices you want to use.

Engine Where it runs Notes
Windows voices (SAPI) Windows The voices installed in Windows, no account needed.
espeak-ng Windows and Linux Free and offline. On Linux install espeak-ng from your distribution (the Flatpak includes it).
ElevenLabs Cloud Needs your ElevenLabs API key.
Azure, Google Cloud, Amazon Polly, IBM Watson Cloud Need your own account and key with that provider.
Acapela, CereProc, TTS Monster, Uberduck Cloud Need your own account and key.

Cloud engines charge per character on the provider's side; Cuebot sends them only the text to speak.

Voice aliases

A voice alias is a name ("Narrator", "Robot") for a voice and its settings (speed, pitch, volume, output device). Speak steps, chat reading, events and scripts use aliases, so you can change the voice behind an alias without editing any action. The default alias is used whenever no alias is given.

Reading chat and events

Under Chat & commands:

  • Read chat aloud, with who is read (everyone, subscribers, moderators...), users who are always or never read, and nicknames spoken instead of usernames.
  • The !tts command, with a viewer cooldown and the aliases viewers may pick. Viewers can set their own nickname if you allow it.

Under Events, choose which follows, subs, gifts, raids, cheers and redemptions are spoken and the text for each (%name%, %gift%, %title%, %input%).

Filters

Filters clean up what viewers write before it is spoken: replace or skip words, links and emotes, limit repeated characters and length. Filter profiles can be applied per alias or to all viewer-written text.

The queue and VoiceGate

Texts are spoken one after another. The Overview page shows what is speaking, what is waiting and what was spoken, with Pause, Skip and Clear. Under VoiceGate & queue you can set a queue limit and what happens when it is full, and turn on VoiceGate: it holds the queue while your microphone is louder than a threshold, so TTS does not talk over you.

Voice control

Audio > Voice Control goes the other way: say a phrase and Cuebot runs an action. Add voice commands, then use them in a Voice Control Command trigger. Recognition runs on your computer (Windows speech recognition, or the Whisper speech-to-text model, downloaded on first use).

Clone this wiki locally