MODULE 06 · INCLUDED IN FOUNDATIONS
The voice: text-to-speech
What this module is
Which text-to-speech engine to use, how to wire it so the avatar and the stream hear the same audio, and the small tricks that make a synthetic voice feel alive.
What you'll build
- A locked-in character voice
- Audio routed through a virtual cable so lip-sync never drifts
- 30–50 pre-rendered stock lines that cover latency gaps
Lessons
- 01Choosing an engine (Kokoro, Chatterbox, ElevenLabs, hosted alternatives)
- 02Wiring it up
- 03Making it sound alive
Checkpoint
The character speaks a reply through the virtual cable, and you can hear it in OBS's audio mixer.