Prompt Each Pixaroma

What it does

Put one prompt in each row and press Run once. The workflow runs again for every prompt, one after another, and all the pictures collect in the same Preview node.

It runs them one at a time, so it uses no more memory than making a single picture. A long list is safe on a small graphics card.

Switching a prompt off

Click a row's ON button and it turns off: the prompt stays where it is but is skipped, so you can try the list without one of them and put it back later without retyping. The counter above always shows how many will actually run.

Copy keeps the switches: a switched-off prompt is copied out as a line starting with #, so pasting that list back in brings the same rows back switched off.

A row is controlled only by its own button, so you can safely start a prompt with a # and number your list. It is only text that arrives in a block, from Paste or from the text input, where a line starting with # is treated as switched off.

How to wire it

• Drag prompt into CLIP Text Encode, where you would normally put your text.
• That is the only wire it needs. Everything after it runs once per prompt on its own.
• index into Save Image keeps the files numbered in the order you typed them.
• The text input on the left adds EXTRA prompts from another node, one per line, run after the rows. Wire Prompt Pixaroma into it and your whole tag library works here.

How it differs from the other prompt nodes

Prompt Each

One Run, one queue entry, one picture per prompt. Everything lands in one Preview.

Prompt Multi, Queue Text

One Run, but the queue panel fills with a separate entry per prompt, so you can cancel them one by one.

Prompt Multi, List Prompts

Hands the whole list to Prompt From List nodes, and you pick each one by number.

Prompt Stack and Prompt Pack

Join pieces of text together into ONE prompt. Those make one picture, not many.

Square brackets make more prompts

Put choices in square brackets and you get all of them:

a [red|blue] car gives two prompts, a red one and a blue one.

Several groups on one line give every combination, so a [red|blue] [car|van] is four prompts. The counter at the top always shows how many you will actually get, so you can see one row turn into six before you press Run.

Curly braces are the opposite and still pick one at random, so {morning|evening} stays one prompt. You can use both on the same line. Write \[ if you want a bracket kept as ordinary text.

The buttons on the node

Add row

Adds an empty prompt at the end and puts the cursor in it.

Clear all

Empties the text in every row, keeping the rows and their switches.

Reset

Back to one empty row.

Copy

Up beside the counter: copies every prompt to the clipboard, one per line.

Paste

Beside it: replaces every row with the clipboard, one prompt per line. This is how a hundred prompts get in from a spreadsheet in one go.

Copy and Paste with long prompts

If any of your prompts runs over several lines, switch Cut pasted and wired text on to Blank line first. On New line the clipboard treats every line as its own prompt, so copying and pasting back would break a long prompt into several rows.

The gear

Opens the settings, described below.

Settings, behind the gear

Cut pasted and wired text on

ONLY affects text that arrives in one block: the Paste button and the text input. Your rows are already separate prompts, so this never changes them. New line makes every line a prompt; Blank line lets a prompt run over several lines and starts the next after an empty line.

Wired prompts go

Before or after your rows. They are always added, never instead, so this is just the order the pictures come out in.

Expand [a|b]

Turn the brackets off if you want them treated as ordinary text.

Trim spaces

Drops spaces at the start and end of every prompt.

Skip empty lines

An empty line is ignored instead of running with no prompt.

Stop after

The most prompts one Run may queue, so a typo cannot start hundreds of renders. Brackets multiply fast: three groups of four options on ONE row is already 64. It counts prompts, not rows.

Using your tag library with it

This node has no tag list of its own, on purpose. Wire Prompt Pixaroma into the text input and your whole @tag library works here: the tags are filled in first, then the result is split into prompts.

Anything on that input is ADDED after your rows, so what is on the node is always part of what runs. To use only the wired prompts, press Reset.

The same trick adds a shared ending to every prompt. Send prompt into Text Join Two Pixaroma with your style words in the second box, and because this node sends a list, the join happens once per prompt.

What comes out

prompt

Each prompt in turn. This is the one you wire to CLIP Text Encode.

index

Which prompt this run is, counting from 1.

total

How many prompts this Run will produce.