Generate audio
Ask the SI Agent (Manager) or use empty-state Generate music. Needs sign-in. Image/audio generations are synchronous; video is queued.
Kinds
| Kind | Notes |
|---|---|
| Music | KnoxChat Music, up to 5 minutes. Can attach a video for video-to-music. |
| TTS | Text to speech. Voice from the role, the selected voice, or the public default Jarnathan. |
| SFX | 0.5–30 s |
Defaults (format MP3 / WAV / Opus, sample rate 16 / 24 / 44.1 / 48 kHz) live in SI Agent Configuration → Audio generation. The agent does not ask you those questions.
Voice pick order (TTS and Voice Changer)
- Explicit
voice_id - Role voice
- Account selected voice (
/v1/knox-studio/selected-voice) - Public default Jarnathan (
c6SfcYrb2t09NHXiT80T). Rachel's id still works for older projects.
Browse voices from Character Roles → Browse Voice Library (My voices, Library, Collections, search).
Where files go
~/.knoxmedia/audio/, then onto the timeline if the agent places them. Score a selected clip from the context menu (Score clip).