Write a character video script with music
Trigger a video project and get an AI script split into timed character lines, plus background music made to fit its length
Version 1
What it does
A webhook starts this for a video project in Airtable. The workflow reads the project, its generation config and the active characters with their roles, voice and image prompts, then builds one full prompt.
An OpenAI model writes the script, B-roll ideas and a music prompt. These are saved as a script generation batch, a script and a music row, and the project status moves on. The script is split into lines with a character, dialogue, order and start and end times. Stable Audio on fal.ai then makes music that matches the script's length, and the link is saved. Last, it calls the speech and lipsync workflow with the script id.