Block Memory

How Memory Works

Every block in Raelume remembers its configuration. Think of each block as a container that stores your settings. When you select a model, adjust parameters, or change settings, those choices are saved automatically and persist across sessions.

This means you can:

• Close your browser and come back later with everything intact
• Build complex workflows where each block uses different models
• Reproduce results because settings are stored with each version

What Gets Saved

Model Selection

Each block remembers which AI model you've chosen. Image blocks remember your Flux or SDXL selection, Video blocks remember your motion model, and so on.

Parameters

All generation parameters are saved per block:

• Resolution and aspect ratio
• Guidance scale and steps
• Seed values (when specified)
• Model-specific settings

Version History

Each version stores a snapshot of the settings used to create it. You can see exactly what model and parameters produced any result.

Auto Switchers

Blocks intelligently auto-switch models based on your inputs:

Image Block

• No image connected → Text-to-image models available
• Image connected → Image-to-image and inpainting models activate
• Multiple images connected → Multi-reference models become available

Video Block

• No image connected → Text-to-video models available
• Image connected → Image-to-video models activate automatically

3D Block

• Single image → Single-view reconstruction models
• Multiple images → Multi-view reconstruction options appear

The block detects your workflow and presents the most relevant model options.

Per-Version Memory

When you generate content, the block stores:

• Which model was used
• All parameter values at generation time
• The prompt that was used
• Input references (what was connected)

This makes every version reproducible. You can always see how something was made and use those settings again.

Pro Tips

1. Each block is independent - use different models in different blocks
2. Check version details to see exactly what settings produced a result
3. Auto-switching helps you discover relevant models for your workflow
4. Settings persist even after browser refresh or closing the tab