Prompt Node Architecture

Prompt nodes are how text instructions enter a generation pipeline. Rather than typing a prompt directly into every AI node, you author it once in a prompt node and wire that node into one or more generation nodes through an edge. This page explains how prompt nodes structure their input and output, and how that output flows into the nodes that consume it.

Inputs and outputs

The structure of a prompt node is intentionally minimal:

• Input— the prompt text itself, which you type into the node's text field. The prompt node has no incoming data ports; its content is authored in place.
• Output — a single text output port (labeled Text) that carries the prompt to downstream nodes.

Because the output is a plain text port, a prompt node can feed anything that accepts a text input — an AI Text node, an AI Image or AI Video node's prompt input, a Text to Speech node, an AI SFX node, and so on. The prompt node also shows how many outgoing connections it currently feeds.

How generation nodes consume a prompt

Generation nodes declare a Prompt input port that accepts text. On the image, video, text, and SFX generation nodes this port is required — the node needs a prompt to run — and it accepts more than one incoming edge, so several text sources can contribute to a single generation.

• AI Text takes a prompt and produces a text output that can itself feed further nodes.
• AI Image and AI Video take a required prompt plus optional image (and, for video, video and audio) reference inputs, and emit a generated image or video.
• Text to Speech and AI SFX take a prompt and produce audio.

Easy Connect: the one-port variant

Easy Connect is a streamlined sibling of the prompt node. It works the same way — you type text, and it exposes a single text output — but it is presented as a quick, low-chrome connector for sending text into a generation node with minimal setup. Use a full prompt node when you want a clearly labeled prompt surface on the canvas, and Easy Connect when you just need to pipe some text along.

Wiring a prompt into a generation

1. 1

   Create the prompt node

   Add a prompt node from the Add node menu, or use the Create prompt node tool in the floating toolbar to type your text and drop the node in one step. See Panels.

2. 2

   Add a generation node

   Add the AI node you want to drive — for example AI Image or AI Text. It will show a required Prompt input port.

3. 3

   Connect output to prompt input

   Drag from the prompt node's Textoutput to the generation node's Prompt input. Because the prompt input accepts multiple connections, you can wire more than one text source into the same generation.

4. 4

   Run

   When the workflow runs, the prompt text flows along the edge into the generation node, which produces its output. Edges on the active path animate during the run.

Related

• Edges — port directions, data types, and connection rules.
• Nodes — the unit of work prompt nodes belong to.
• Node Types — the reference for each generation node a prompt can feed.