Save 4 hours! We auto-setup your workflow! Free!

Drop your workflow.json — we handle every dependency, custom node, and model. Just open the link and run.

Auto-Setup Workflow Json (Free) Now!
ComfyUI > Nodes > FunUI > FunUI Prompt Markup

ComfyUI Node: FunUI Prompt Markup

Class Name

FunUIPromptMarkup

Category
FunUI/Conditioning
Author
digital-garbage (Account age: 1276days)
Extension
FunUI
Latest Updated
2026-11-03
Github Stars
0.02K

How to Install FunUI

Install this extension via the ComfyUI Manager by searching for FunUI
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter FunUI in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

FunUI Prompt Markup Description

Separates clean prompt text from weight, timing, and blend markup for later application.

FunUI Prompt Markup:

FunUIPromptMarkup prepares prompts containing special markup so that your text encoder can read the intended description without treating the markup symbols as ordinary punctuation. It recognizes three documented forms: weighted wording such as (word:1.5), timed phrases such as [phrase@2-4], and blends such as [a|b]. It separates the clean prompt text from the instructions that need to be applied later in the workflow. This separation is especially important for encoders such as Qwen, which otherwise read weighting syntax as part of the sentence rather than as a generation control. The node handles preparation only: it does not encode text, modify a model, or apply these effects by itself. Together with the companion FunUI Apply Prompt Markup node, it provides a way to preserve readable prompt text while carrying markup instructions forward for supported models.

FunUI Prompt Markup Input Parameters:

text

A required STRING containing the prompt you want to prepare. The field supports multiple lines and has an empty-string default, ""; no minimum or maximum length is specified in the node interface. You can enter ordinary descriptive text alongside documented markup forms such as (word:1.5) for weighting, [phrase@2-4] for a timed phrase, and [a|b] for a blend. The node parses this text, separates recognized markup from the text intended for encoding, and counts the recognized instructions by category. For example, you could start with A portrait of a (dancer:1.5) to prepare a weighted instruction while keeping the encoder-facing description free of the weighting notation. This node exposes no separate weight, timing, or blending controls; those instructions are written directly in the prompt. The supplied source does not specify numerical limits for values inside markup or the handling of every malformed syntax case, so use the documented forms rather than assuming additional syntax is supported.

FunUI Prompt Markup Output Parameters:

clean

A STRING containing the parser’s cleaned prompt text, intended for your text-encoding node. Its purpose is to keep recognized markup notation out of the sentence the encoder reads, while the associated instructions are carried separately. Connect this output to the encoder that produces the positive conditioning you will send to FunUI Apply Prompt Markup. Keeping that connection consistent matters because the companion node needs to locate the marked words within the encoded version of this same text. Avoid rewriting the cleaned text between parsing and encoding, as changes can prevent the companion node from matching the instructions to the conditioning.

markup

A FUNUI_MARKUP value containing the structured parsing result, including the cleaned text and the recognized weighted, timed, and blended instruction collections. This is workflow data rather than a finished prompt or an encoded conditioning value. Connect it to the companion FunUI Apply Prompt Markup node so that the instructions can be interpreted after text encoding, alongside the model, encoder, conditioning, and other required information. Producing this output does not mean the effects have already been applied: successful application depends on the downstream node and on model-module support for locating words and, when needed, mapping video time to model rows.

status

A STRING summarizing how many instructions the parser recognized in each category. Only categories with nonzero counts are included, in the order weighted, timed, then blended; for example, a result may read 1 weighted, 1 timed. If none of these collections contains instructions, the value is exactly no markup. Use this output to check whether your prompt produced the expected categories of markup. It reports parsing counts only, not whether a model supports the requested effects or whether the companion application node successfully applied them.

FunUI Prompt Markup Usage Tips:

  • Connect clean to your text encoder, then send the resulting positive conditioning and the matching markup value to FunUI Apply Prompt Markup. The intended sequence is prompt preparation, text encoding, and finally markup application.
  • Start with one documented instruction, such as A portrait of a (dancer:1.5), and check that status reports 1 weighted before building a more complex prompt. This makes it easier to identify which edit caused an unexpected result.
  • Keep the cleaned prompt unchanged before encoding. If you edit or rewrite it, run the revised prompt through FunUIPromptMarkup again so that the text and markup remain synchronized.
  • Treat [phrase@2-4] as timing information whose application depends on the model’s video-time support. The module describes timing in seconds, but this preparation node does not itself translate those seconds into video positions.
  • Do not rely on an enabled FunUI Enhance Prompt node to preserve markup during rewriting. The enhancer explicitly warns that markup does not survive a rewrite; finish your rewriting first, then add or verify the markup before passing the final text through FunUIPromptMarkup.
  • Check the companion application node’s status as well as this node’s status. Recognized markup can still be refused downstream if the encoder or encoded prompt cannot be matched.
  • Account for the attention implementation when planning performance. The module states that weights use an attention bias and that SLA runs biased calls with dense attention, so weights or timed phrases can change the attention execution path; this preparation node has no setting to override that behavior.

FunUI Prompt Markup Common Errors and Solutions:

no markup

  • Explanation: This is an informational status, not an error. The parser returned no weighted, timed, or blended instructions. That is expected for an ordinary prompt, but it can also indicate that the notation you intended was not recognized.
  • Solution: If you want markup effects, test a documented form such as (word:1.5), [phrase@2-4], or [a|b] and inspect the status again. If you only want ordinary text, no correction is necessary.

not applied: this text encoder is not one a model module can place words in

  • Explanation: This message comes from FunUI Apply Prompt Markup, not from the preparation node. The companion node could not obtain a supported tokenizer through the model-module capability system, so it could not locate the marked wording in the encoded prompt. The prompt still proceeds without the markup effects.
  • Solution: Use an encoder and model combination supported by the installed FunUI model modules. Verify compatibility rather than assuming that every text encoder can apply markup; recognizing the notation in FunUIPromptMarkup alone does not establish downstream support.

not applied: the encoded prompt does not match this text -- was the clean text from Prompt Markup what got encoded?

  • Explanation: This downstream message means the companion node could not obtain a valid mapping between the cleaned prompt and the encoded conditioning. A likely workflow issue is encoding the original marked prompt, changing the text after parsing, or pairing conditioning with markup from a different prompt.
  • Solution: Connect clean directly to the encoder and send that encoder’s positive conditioning to the companion application node together with markup from the same FunUIPromptMarkup execution. Remove intervening text rewrites and regenerate both branches from the same prompt.

the prompt has (word:1.5) / [a@2-4] / [a|b] markup, which the model will read as plain text: with the enhancer on, markup does not survive a rewrite

  • Explanation: This warning comes from FunUI Enhance Prompt when it detects markup in text being rewritten. The rewrite process does not preserve the markup instructions reliably, so the result may no longer contain the controls you intended.
  • Solution: Disable enhancement for an already marked prompt, or finish enhancement first and add the markup afterward. Pass the final marked text through FunUIPromptMarkup before encoding.

Markup is recognized, but no effect is visible

  • Explanation: This is a troubleshooting symptom rather than a literal error message in the supplied code. FunUIPromptMarkup only separates text and instructions; a status such as 1 weighted confirms recognition, not application. Effects will not be applied merely by encoding clean.
  • Solution: Verify that markup is connected to FunUI Apply Prompt Markup and that the companion node receives the corresponding encoded conditioning and required model-related inputs. Check its status for refusal messages, and confirm that your downstream generation path uses its returned model and conditioning.

The encoder reads markup symbols as part of the prompt

  • Explanation: This is a workflow symptom rather than a literal error message. It can occur when the original prompt containing notation is sent directly to the encoder instead of the cleaned text. The module specifically notes that Qwen reads notation such as (word:1.5) as punctuation in the sentence.
  • Solution: Replace the encoder’s original prompt connection with the clean output from FunUIPromptMarkup. Carry the instructions separately through markup to the companion application node.

FunUI Prompt Markup Related Nodes

Go back to the extension to check out more related nodes.
FunUI
RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

FunUI Prompt Markup