Design a custom voice for each character

This is the third step of the character video system. The image workflow calls its webhook when every portrait is done, or you can start it from an Airtable button.

It marks the Generate Voices Status as in progress and loads every character linked to the generation config. For each character it sends the voice prompt to ElevenLabs voice design, then creates the voice from the preview. The voice id, settings and sample link are saved on the character and in a character voices table. When the loop ends it marks voices complete and calls the script workflow's webhook so the next step can start.