Negative Wildcard Processor:
This node helps you keep prompt text organized by moving specially marked phrases from the positive prompt into the negative prompt. It also removes LoRA-style tags from the positive text. This is useful when a prompt contains instructions that should be excluded rather than encouraged. The node edits text only; it does not generate an image, apply a LoRA, or interpret the meaning of the phrases it moves.
Negative Wildcard Processor Input Parameters:
positive
A required STRING containing the prompt text to process. Place a phrase between <! and !> to have the node remove that marked section from the positive text and append the phrase inside it to the negative text. For example, portrait <!blurry!> becomes portrait on the positive side, while blurry is added to the negative side. The node also removes text matching a <lora...> tag from this input, without moving that tag to the negative text. The source defines an empty-string default, multiline text, and forceInput=True, so this input is intended to receive a connection. No minimum or maximum text length or selectable options are specified.
negative
A required STRING containing the starting negative prompt. The node preserves this text and appends each phrase extracted from positive, in the order found, with a space before each phrase. If there are no marked phrases, it returns this text unchanged. The source defines an empty-string default, multiline text, and forceInput=True, so this input is intended to receive a connection. No minimum or maximum text length or selectable options are specified.
Negative Wildcard Processor Output Parameters:
text_positive
A STRING containing the processed positive prompt. Sections matching <!...!> and tags matching <lora...> are removed; other text remains in place. Removal does not tidy surrounding spaces or punctuation, so you may see extra whitespace where a section was deleted.
text_negative
A STRING containing the original negative prompt followed by the text from each matched <!...!> section. The markers themselves are not included. Each extracted phrase is appended with a leading space, in the order it appeared in positive; the node does not add commas, remove duplicates, or check whether a phrase was already present.
Negative Wildcard Processor Usage Tips:
- Put only the phrase you want moved inside each
<!...!>section. For example,portrait <!blurry!> <!watermark!>appendsblurryandwatermarkto the negative prompt. - Include your own punctuation if your downstream prompt expects comma-separated terms. The node adds spaces before extracted phrases, not commas.
- Connect the processed outputs to the appropriate downstream positive and negative prompt inputs. If you need LoRA tags to remain in the positive text, do not pass them through this node.
- Check the resulting text when spacing matters: removing marked sections or LoRA-style tags can leave gaps, and an empty starting negative prompt produces a leading space when a phrase is appended.
Negative Wildcard Processor Common Errors and Solutions:
No phrase was moved to the negative prompt
- Explanation: The node only extracts text enclosed by the exact markers
<!and!>in the positive input. Plain negative-sounding words are not moved automatically. - Solution: Check that the phrase is in
positiveand written as, for example,<!blurry!>.
Part of a marked phrase remains in the positive prompt
- Explanation: The matching pattern does not span a line break inside
<!...!>. An unfinished or mismatched marker also will not match. - Solution: Keep each marked phrase on one line and check both markers.
A LoRA-style tag disappeared
- Explanation: The node deliberately removes text matching
<lora...>from the positive output; it does not transfer that text to the negative output. - Solution: If the tag must be retained, place or apply it elsewhere in your workflow rather than relying on this node’s positive output to carry it forward.
The output has extra spaces or unexpected separators
- Explanation: Removed sections leave surrounding text untouched, and extracted phrases are appended using a single leading space without adding commas.
- Solution: Adjust punctuation in the input or clean up the resulting text in a later step if your prompt format requires it.
The node reports a missing or invalid input connection
- Explanation: Both inputs are required STRING inputs configured with
forceInput=True. The exact validation message depends on the application running the node. - Solution: Connect a source that supplies STRING text to both
positiveandnegative, even when one prompt should start empty.
