Raelume
RaelumeDocs
Back to Blocks
Core Concepts

Context

Understand how data flows between blocks and how to pass content from one to another.

What is Context?

Context is the data that flows between blocks when they're connected. When you draw a connection from one block to another, you're passing context - images, text, videos, or 3D models that the receiving block can use.

Think of it like passing ingredients between stations in a kitchen. The output of one station becomes the input for the next.

How Context Flows

When you connect blocks:

  1. The source block produces output (an image, text, video, etc.)
  2. The connection carries that output to the next block
  3. The target block receives it as input and can use it
flowchart LR
    T[Prompt Agent Block]
    I[Image Block]
    V[Video Block]

    T -->|text| I
    I -->|image| V

The currently selected version in a block is what gets passed downstream. Change your selection, and all connected blocks receive the new content.

Context Types

Different blocks produce and accept different types of context:

Image Context

  • Produced by: Image, Variator
  • Accepted by: Video, 3D, Image, Prompt Agent, Webhook
  • Use: Visual reference for generation, starting frame for video, source for 3D conversion

Prompt/Text Context

  • Produced by: Prompt Agent, Prompt
  • Accepted by: Image, Video, 3D, Audio, Worlds, Webhook
  • Use: Prompts, instructions, structured data

Video Context

  • Produced by: Video, Clip Assembler
  • Accepted by: Clip Assembler, Webhook
  • Use: Clips for timeline editing, content for export

3D Context

  • Produced by: 3D
  • Accepted by: Webhook
  • Use: GLB models for download or integration

Multi-Input Context

Some blocks accept multiple context inputs:

  • Image blocks can receive both images (for image-to-image) and text (for prompts)
  • Video blocks can receive images (starting frame) and text (motion description)
flowchart LR
    T[Prompt Block]
    R[Reference Image]
    I[Image Block]

    T -->|text| I
    R -->|image| I

When multiple inputs are connected, blocks intelligently combine them - using images as visual reference and text as generation guidance.

Variator Pass-Through

The Variator block works as a context pass-through that branches your workflow:

  1. Accepts one media input (image, video, or 3D) plus optional text context
  2. Generates N variations based on the number of variation slots you add
  3. Each variation output passes both the original media AND the variation text to connected blocks
flowchart LR
    T[Prompt Agent]
    S[Source Image]
    VAR[Variator]
    I1[Image Block 1]
    I2[Image Block 2]
    V1[Video Block]

    T --> VAR
    S --> VAR
    VAR -->|img + var 1| I1
    VAR -->|img + var 2| I2
    VAR -->|img + var 3| V1

This means downstream blocks receive full context - the visual reference plus the variation-specific instructions - enabling parallel generation workflows from a single source.

Pro Tips

  1. Connect a Prompt block to multiple target blocks to reuse the same prompt across all of them
  2. Use Variator to branch one image into multiple parallel workflows with different variation prompts
  3. Check the block's documentation for supported input types
  4. The version indicator shows which output is currently being passed downstream
VersionsMemory