​

Continuity Enhance

The Continuity Enhance node analyzes your scenes and generates reference descriptions for characters and backgrounds, ensuring visual consistency when generating images or videos across multiple shots.

​

Inputs

​

Outputs

​

How it works

1. Receives an array of scene descriptions (typically from a Parse Scenes node)
2. Analyzes the scenes to identify recurring characters and backgrounds
3. Generates detailed reference descriptions for each character and background
4. Injects these references into downstream image/video generation prompts

In parallel mode (within a Split loop), the first iteration’s generated image is used as a visual reference for subsequent iterations, maintaining consistency. In sequential mode, continuity references are passed between iterations so each shot builds on the previous one.

​

Credit cost

Paid — cost depends on the text model used for analysis (typically 1 credit with Gemini Flash Lite).

​

Configuration

This node runs as a background step by default. It processes automatically without user interaction.

​

Tips

• Place this between Parse Scenes and your image/video generation nodes
• Works best with a Split → Collect loop for multi-shot workflows
• The style prompt input lets you define an overall visual style that applies to all shots
• Combine with Multi-Angle Generation for consistent character views from different angles

​

Example use cases

• Ensuring the same character appears consistently across all shots in a multi-scene video
• Maintaining a consistent background style (e.g., “cyberpunk city”) across different scenes
• Creating coherent multi-shot narratives where characters and locations stay recognizable