GUIDE

Twitch chat TTS with character voices

Chat text to speech turns viewer messages into part of the show: a donation read by a villain, a channel point redeem in a cartoon voice, a greeting from the whole chat. It is also the feature most likely to go wrong if anyone can make the stream say anything.

SuperClonica voices chat messages on your own streaming PC with natural and character voices, in well under a second per message. This guide shows how chat TTS works with it, which rules keep it fun, and how to set it up.

Browse voices Install the SuperClonica engine

How chat TTS works

SuperClonica does not plug into Twitch by itself: the bot is your own code or a tool that can send an HTTP request.

  • A bot reads new chat messages, or only redeems, bits and highlighted messages.
  • It filters and shortens each message.
  • It sends the text to a local SuperClonica server and plays the WAV that comes back on stream.
  • A short message takes well under a second on an ordinary processor.

Pick the chat voice free

On the page of any voice, paste a typical chat message, 30 to 100 characters, choose the language and press Voice it. In a few seconds you hear how your chat will sound, with no account.

Rules that keep chat TTS fun

Chat shorthand is spelled out by the engine: idk becomes “I don’t know”, glhf becomes “good luck have fun” and brb becomes “be right back”, and emoji are cleaned up instead of read as symbols.

  • Voice only redeems, bits or subs: it keeps spam off the air.
  • Cap the length, for example 200 characters.
  • Filter links, slurs and spam with your block list and your platform’s moderation tools.
  • Queue messages so they never overlap, and give moderators a skip key.

Set it up

Start the server on your streaming PC: docker run -d -p 8000:8000 -v superclonica-models:/models superclonica/superclonica:cpu. Your bot sends each message as JSON to POST http://localhost:8000/synthesize, for example {"text": "glhf, see you after the raid", "voice": "F1", "lang": "en"}, and plays the WAV through an audio output that OBS captures. The bot can pick a voice per viewer or per redeem; for a catalog voice, send the content of its voice file in the voice field. Set SUPERCLONICA_LANES to voice several messages at once.

Voices and rights on stream

Character voices are fan-made and not endorsed by the rights holders; they are fine for fun on stream within the SuperClonica terms. Never let chat speak with a voice that imitates a real person: anyone could make that person “say” anything. Follow your platform’s rules on AI-generated audio.