Alerts and text-to-speech

Text-to-speech voices

Choose the voice that reads donation and sub messages, use your computer's voices or your own ElevenLabs or Fish Audio key, and tune rate, pitch and volume.

Updated 2026-10-16 · support@konduit.gg

Related help

Konduit can read donation and sub messages out loud. You choose the voice under TTS Provider on the Alerts & TTS page. The voice always plays from the Konduit app, not from the Browser Source in OBS.

Use your computer's voices

  1. Open Alerts & TTS.
  2. Under TTS Provider, choose Web Speech API. It uses the system voices from your computer, with no cost and no setup.
  3. Pick a Voice and choose Preview voice to hear it.
  4. Set Rate, Pitch and Vol.

Use your own ElevenLabs or Fish Audio voice

You need an account and an API key from ElevenLabs or Fish Audio. Get the key from that service's own website.

  1. Open Alerts & TTS.
  2. Under TTS Provider, choose Bring Your Own Key.
  3. In Provider, choose ElevenLabs or Fish Audio.
  4. Paste your key into API Key and click away from the field. Konduit shows Saved and loads your voices. To remove the key, empty the field and click away; Konduit shows Removed.
  5. Pick a Voice. If a voice you just made is missing, choose Refresh next to Voice.
  6. Choose Preview voice to hear it.
  7. Pick a Backup system voice. If your voice service can't read a line (no credits, no connection, or a key problem), Konduit reads it with this voice instead.

Fish Audio extras

  • Model picks the Fish Audio model: Drama 3 (preview) (the default), S2.1 Pro or S2.1 Pro Free.
  • To use a voice from Fish Audio's public library, paste its fish.audio voice link under Add a voice from the Fish Audio library and choose Add. Save your Fish Audio key first. A voice you added can be taken off the list with Remove this voice.

Tune the voice

Rate, Pitch and Vol shape both voices: your own ElevenLabs or Fish Audio voice and the backup system voice.

Turn reading on or off

Further down the same page, each lane has its own switch:

  • Enable donation TTS reads donation messages. It only reads when the donor wrote a message.
  • Enable sub TTS reads sub messages. Subs are read only when the subscriber typed a message.

Donation TTS interval and Sub TTS interval set the pause between reads. Under Variable Substitution you choose how a donation is announced, for example with or without the donor's name and amount. Platform TTS Thresholds set the amount below which an alert is shown without being read.

If it doesn't work

  • Nothing is read: check that Enable donation TTS or Enable sub TTS is on, and that the message wasn't below the amount in Platform TTS Thresholds.
  • Your own voice doesn't play but the backup voice does: your ElevenLabs or Fish Audio key was refused or the account has no credits. Check the key and your account with that service, then paste the key again.
  • The voice list is empty: check that the key in API Key is saved, then choose Refresh.