AI character video regeneration on Airtable
Redo any single part of an AI talking-character video: voice, image, script, speech or lip-sync clip, tracked in Airtable
Version 1
What it does
This holds seven webhooks, one per redo job, for an AI character video project stored in Airtable. You can regenerate a character's voice with ElevenLabs voice design, or its image with fal (from a training image when the background stays the same, or from text).
You can also regenerate a full script: an OpenAI model writes new lines for your characters, background music is made to the script's length, and every line is voiced and turned into a lip-synced video clip. Single lines can be revoiced with ElevenLabs, uploaded to Cloudinary and re-animated. Each job updates its status field, and the steps call each other's webhooks.