​

Parse Scenes

The Parse Scenes node takes a block of text (typically a script) and splits it into individual scene descriptions. Each scene becomes an item in a text array that can be fed into a Split node for per-scene processing.

​

Inputs

​

Outputs

​

How it works

1. Local parsing — If the input text already has numbered lines (e.g., 1. First shot\n2. Second shot), scenes are extracted instantly without using AI credits.
2. AI parsing — If the text is free-form (no numbered structure), the node uses AI to decompose it into a numbered scene list. This costs 1 credit.
3. Inline editing — Click any scene to edit its description before confirming.

​

When to use

• Before a Split node to process each scene individually through Image Gen, Video Gen, Trim, etc.
• As a standalone scene parser in composable multi-shot workflows.
• Whenever you need to decompose a script into discrete visual shots.

​

Tips

• Connect a Script Generation node before this to get properly formatted numbered scripts (saves credits since no AI parsing is needed).
• The node preserves scene order, which is important for sequential Split mode (visual continuity).