Motion graphics for YouTube from a transcript
Pick a transcript in Airtable and get still images from Nano Banana, animated with VEO into short clips saved back to Airtable
Version 1
What it does
A webhook starts it from a button on an Airtable record holding a cleaned transcript, the number of clips you want and a linked brand with its style and color palette. It sets the record's AI status to AI Writing.
An OpenAI model reads the transcript and plans short motion-graphic sections of up to eight seconds, each with timestamps, a type, an image prompt and an animation prompt. For each section it makes a still with Nano Banana on fal.ai, then animates it into an eight second 1080p clip with VEO 3 Fast. Each clip and image is saved as a new record linked to the transcript, and the prompts are written back to the source record, marked AI Complete.