Write YouTube titles and descriptions for a session

A webhook starts it from a session record in Airtable. It marks the session's video details status as AI Writing and loads the video prompts linked from your prompt library.

If the session already has a cleaned transcript, an OpenAI model uses it. If not, the workflow downloads the session's VTT caption file and turns it into a timestamped transcript first. For each prompt, the model writes the YouTube upload assets (titles, descriptions and similar) from your brand context, the summit and speaker details and the transcript. Each result is saved to your content values table as a To Do row of type Session Video. Usage is logged and the status changes to AI Complete.