Creating Pipelines

A pipeline is a chain of connected nodes that produces a final output. This tutorial walks you through building one from scratch.

Step 1: Open Spaces

1. Click Spaces in the left sidebar
2. Click Create New to start with a blank canvas, or open an existing space

Step 2: Add Nodes

Add nodes to the canvas using any of these methods:

• Click the Add Node button in the toolbar and select a type
• Use keyboard shortcuts: I (Image), V (Video), M (Music), T (Speech)
• Right-click the canvas for the context menu

Position nodes left-to-right to keep your pipeline readable. Upstream nodes (inputs) go on the left, downstream nodes (outputs) go on the right.

Step 3: Connect Nodes

1. Hover over the output port on the right side of a node
2. Click and drag to the input port on the left side of the target node
3. Release to create the connection

Connections determine data flow. When an Image node is connected downstream from a Prompt node, it automatically receives the prompt text.

Valid Connection Examples

• Character + Scene + Shot → Image — Combines character, environment, and framing into one image generation
• Image → Video — Animates a generated image into a video clip
• Video + Music → Merger — Combines a video with a music track
• Speech + Video → Lipsync — Syncs generated speech with a character video

Step 4: Configure Parameters

Click any node to open its settings panel on the right side. Common parameters include:

• Model — Which AI model to use for generation
• Aspect Ratio — Output dimensions (1:1, 16:9, 9:16, etc.)
• Quality — Generation quality level
• Prompt — Text description (on Prompt nodes)
• Duration — Length in seconds (on Video and Audio nodes)

Step 5: Run the Pipeline

1. Click the Run button in the toolbar
2. Nodes process in order, starting from the leftmost upstream nodes
3. A progress indicator appears on each node as it generates
4. Generated assets appear directly on the nodes when complete

Tip: You can also run individual nodes by clicking the play button on a specific node. This is useful for regenerating just one step without re-running the whole pipeline.

Understanding Data Flow

Data flows upstream to downstream through connections. A downstream node automatically inherits outputs from all connected upstream nodes. For example:

• A Prompt node sends its text to a connected Image node
• The Image node sends its generated image to a connected Video node
• The Video node sends its clip to a connected Merger node

This chaining means you can build complex multi-step workflows without manually copying outputs between steps.

Next Steps

• Publishing from Spaces
• Spaces Overview