Workflow operators

Operators work on the text before it reaches the Generator: Merge, Caption, Style and Pose.

Merge joins incoming prompts into one text, in letter order; you set the number of inputs in its menu, from two to twelve.

Caption turns an image into text. We wire the taco truck photo into Caption, and its output into Merge's second input.

Caption is billed per image; its estimated five credits are added to the total in the run bar.

Double-click a Style node to open Prompt Lab: one thousand and sixty presets across style, camera, light, lens, stock and palette.

Under Light, we search for "neon" and pick Neon Wash. The preset's description is added to the end of the incoming text, and the node takes on its artwork.

You can chain Style nodes: in the second one, we pick the 500T negative from Stock.

A Pose node opens the Pose Atlas: three hundred and fifty poses in four classes. The pose you pick is added to the prompt as a body instruction.

From Fashion & Advertising, under Commercial & Product, we pick Two-Hand Display.

When you run it, the operators build the text in order: the caption cost three credits and the image one hundred.

Caption's menu shows the description it wrote, and in the result the scene from the photo, the cook and the neon light come together in one image.

Click the image once and it opens full size in a new tab.

In short: Merge gathers texts, Caption turns an image into text, and Style and Pose add to it. In the last part, we'll cover managing your runs.