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 > ComfyUI-ialhabbal > Prompt Verify

ComfyUI Node: Prompt Verify

Class Name

Prompt Verify

Category
text
Author
ialhabbal (Account age: 0days)
Extension
ComfyUI-ialhabbal
Latest Updated
2026-10-29
Github Stars
0.01K

How to Install ComfyUI-ialhabbal

Install this extension via the ComfyUI Manager by searching for ComfyUI-ialhabbal
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter ComfyUI-ialhabbal 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

Prompt Verify Description

Reviews or edits prompts during a workflow, then outputs text and optional CLIP conditioning.

Prompt Verify:

Prompt Verify gives you a chance to review or revise prompt text while a workflow runs. It sends the current text to the interface, waits for a submitted response, and then passes the chosen text onward. This can help when you want to inspect an automatically generated prompt before it reaches the next stage of your workflow. It can also pass text you write directly in the node without opening the review step.

Prompt Verify Input Parameters:

use_external_text_input

A BOOLEAN switch that controls whether the connected text value is considered. It defaults to False; its options are on and off. When off, the node ignores text even if it is connected. Turn it on when you want to review a prompt supplied by another node. If both external text and a nonempty LLM input are enabled, the LLM input takes precedence.

use_llm_input

A BOOLEAN switch that controls whether the connected llm_input is requested and can supply the prompt for review. It defaults to False; its options are on and off. When on, a nonempty llm_input replaces the selected text value. If the LLM input is empty, it does not replace that value, so check the upstream result when the prompt being reviewed is not what you expected.

timeout

An INT setting for how long the node waits for a review submission before notifying the interface that time is up. The default is 60 seconds, the minimum is 1, the maximum is 2400, and the adjustment step is 1 second. After sending the time-up notification, the node waits up to five additional seconds for a final submission. A shorter timeout reduces waiting when you do not expect to edit; a longer one gives you more time to read and revise a detailed prompt.

clip

An optional CLIP connection used to turn the final text into conditioning for image-generation steps that accept it. No default, minimum, or maximum is specified. If clip is not connected, the node still returns the final text, but COND is None. If encoding fails, the text is still returned and COND is also None.

text

An optional STRING connection for prompt text from another node. It has no specified default, minimum, or maximum and must be supplied through a connection rather than typed into this input. It is considered only when use_external_text_input is on. If use_llm_input is also on and supplies nonempty text, that LLM text replaces it.

llm_input

An optional STRING connection intended for text produced by an LLM. It has no specified default, minimum, or maximum and must be supplied through a connection. Turn on use_llm_input to use a nonempty value from this connection in place of the selected text. An empty LLM result does not override the selected text.

editor

An optional, multiline STRING field where you can type prompt text directly. Its default is an empty string; no minimum or maximum length is specified. The field’s tooltip says to press Shift+Return to submit an edit. When neither external nor LLM input provides selected text and editor is nonempty, the node returns the editor text directly, without entering its timed review loop. Otherwise, it proceeds with the review step, using the selected input text.

Prompt Verify Output Parameters:

COND

The CONDITIONING output produced by encoding the final prompt with the connected clip. Connect it to a downstream step that expects conditioning. Its value is None when no CLIP model is connected or when CLIP encoding fails, so a populated TEXT output does not guarantee populated conditioning.

TEXT

The STRING output containing the final prompt text. It contains the directly returned editor text when that path applies. During timed review, it contains a nonempty submitted message if one arrives; otherwise, it falls back to the text selected before review. If no source supplies text, that fallback is an empty string.

Prompt Verify Usage Tips:

  • To review text from another node, connect it to text and turn use_external_text_input on. Connect a CLIP model as well when the next step needs COND.
  • To review LLM-generated text, connect it to llm_input and turn use_llm_input on. Remember that a nonempty LLM result takes precedence over text when both switches are on.
  • For a prompt you have already written in editor, leave the external and LLM inputs unused. A nonempty editor value then passes through immediately, which avoids an unnecessary wait.
  • Set timeout to allow enough time for the kind of review you are doing. For example, a brief confirmation can use a shorter setting than a substantial rewrite; account for the additional five-second grace period after time-up.

Prompt Verify Common Errors and Solutions:

[PromptVerify] CLIP encoding failed: ...

  • Explanation: The node prints this message when its attempt to tokenize or encode the final text with the connected CLIP model raises an error. TEXT is still returned, but COND becomes None. The detail after the colon depends on the underlying failure.
  • Solution: Check that the clip connection supplies a working CLIP model compatible with the text you are processing. Review the specific error detail in the log, then rerun the workflow after correcting the CLIP connection or upstream model issue.

COND is empty or None

  • Explanation: This is an output state, not a specific error message from the node. It occurs when clip is not connected or CLIP encoding fails.
  • Solution: Connect a CLIP model if your downstream step requires conditioning. If one is already connected, look for the [PromptVerify] CLIP encoding failed: ... message to identify the cause.

The review opens with an empty prompt

  • Explanation: No source supplied selected text. For example, text is ignored while use_external_text_input is off, and an empty llm_input does not provide replacement text. The node then starts review with an empty string unless a nonempty editor value qualifies for direct return.
  • Solution: Check that the intended connection contains text and that its corresponding switch is on. To pass typed text directly instead, enter it in editor without supplying selected external or LLM text.

The reviewed prompt is not the connected text

  • Explanation: A nonempty llm_input replaces text when use_llm_input is on. Alternatively, text is ignored when use_external_text_input is off.
  • Solution: Turn off use_llm_input if you want the external text to take priority, and turn on use_external_text_input to enable that connection.

The node waits longer than timeout

  • Explanation: Reaching timeout sends a time-up notification, but the node then allows up to five more seconds for a final submission. This extra wait is expected behavior.
  • Solution: Include the grace period when choosing a timeout. Lower timeout if you need the workflow to move on sooner, within the permitted range of 1 to 2400 seconds.

A blank submission does not replace the original prompt

  • Explanation: During timed review, the node uses the submitted message only when it is nonempty; an empty message falls back to the selected input text.
  • Solution: Submit nonempty replacement text when you need a change. Do not rely on a blank submission to clear an existing prompt.

Prompt Verify Related Nodes

Go back to the extension to check out more related nodes.
ComfyUI-ialhabbal
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.

Prompt Verify