Practical guide

Set up chat text-to-speech

Choose who can speak, assign voices and follow a message through the audio queue.

What you will be able to doEligible messages use your chosen voices and you know how to investigate a silent or skipped message.

Before you start

  • Access to TTS in your invited account.
  • An intended audience and an audio route that your stream captures.

Decide who can speak

Choose command-triggered speech or the supported comment mode. Configure audience eligibility, length limits and cooldowns before leaving speech enabled for a live chat. Use the actual trigger shown in TTS settings; a command from another tool is not automatically a TTS command.

Follower, subscriber, moderator, Heart Me and other eligibility depend on available identity data. A Heart Me minimum needs a reported membership level. Review the special rules shown for verified moderators and Super Fans instead of assuming every restriction treats every role identically.

Choose and assign voices

  1. Choose an available voice and audition it locally.
  2. Use voice assignment controls if a particular viewer or group should have a different voice.
  3. Review pronunciation and expressive options for the selected provider. Not every voice engine supports the same controls or languages.
  4. Enable the intended mode and inspect a message in the decision or queue history. Accepted text, generated audio and completed playback are different stages.

Understand generation and playback

Voice generation can use credits. Device speech, cached playback and provider-generated speech have different rules. Read the current credit explanation and pricing before choosing an engine for a busy chat; do not assume that every available voice is unlimited or free.

Local speech has device, browser, memory and language requirements. A downloadable desktop build does not make every speech engine work identically on every operating system.

When speech does not play

  • Not accepted: check audience, command, length, bans and cooldown in the decision history.
  • Accepted but not generated: check provider availability, credits and the error shown.
  • Generated but silent: check mute, queue, active output and playback permission.
  • Audible locally but not on stream: check OBS or LIVE Studio capture, mixer and the selected source.

These instructions describe the reviewed iTok beta. Provider access and physical-device behavior still need to be checked in your setup.

Read product updates