How the script is structured
The script is divided into scenes. Each scene corresponds to a section of your Velo and has two parts:- Visual - describes what the agent will focus on or interact with on the page during that scene
- Narration - the voiceover text the VeloTwin will speak
Navigating scenes
Use the ← → arrow buttons to move between scenes. The current scene number and total count are shown (e.g., SCENE 1) along with a timestamp range (e.g., 00:00 - 00:08) showing where in the demo that scene falls.Editing narration
The Narration text for each scene is directly editable - click in the text area to make changes. If you want to revert your edits and restore the AI’s original text, click Reset next to the narration field.Giving feedback
Each scene has a thumbs up / thumbs down control. Use this to flag scenes that are on the right track or need improvement. This feedback helps guide Velo if you choose to regenerate the script.Regenerating the script
If the overall script needs a more substantial rework, click Regenerate Script at the bottom of the page. Velo will rewrite the script from your narration transcript. You can regenerate as many times as needed.The Visual direction for each scene is set by the agent and is not directly editable. If the agent is focusing on the wrong part of the page, try adjusting your Additional Instructions in VeloTwin Setup before regenerating.