ComfyUI-Przewodo-Utils Introduction
ComfyUI-Przewodo-Utils is a collection of custom nodes for ComfyUI. It brings common workflow tasks, such as comparing values, choosing between inputs, sizing images, and handling lists, into dedicated nodes. It also includes tools intended for video workflows using Wan2.1.
If you build workflows by connecting many small steps, these utilities can make the canvas easier to follow. For example, you can choose a video generation mode in one place, calculate image dimensions while preserving an aspect ratio, or check the shape of a latent when a workflow produces an unexpected result.
The nodes are grouped into two ComfyUI categories:
- PrzewodoUtils/Wan: Tools for Wan-focused video generation and processing.
- PrzewodoUtils: General-purpose tools for images, logic, lists, and workflow checks.
How ComfyUI-Przewodo-Utils Works
A ComfyUI workflow passes information from one node to another. Think of each node as a step at an artist’s workbench: one prepares an image, another makes a decision, and another passes the result to the next stage. This extension adds steps that would otherwise take several nodes to assemble.
Some nodes make choices. For instance, a comparison node can check two numbers and return a different string depending on the result. Other nodes prepare data, such as calculating a resize factor or building a list of LoRAs. The debugging node helps you inspect data when a workflow is difficult to troubleshoot.
The Wan nodes serve more specialized video workflows. They help select generation settings, manage multiple LoRAs, work out suitable image dimensions, and process video data. The extension also describes support for multi-chunk video workflows and performance features. Your available options and results may depend on the rest of your ComfyUI workflow and the models you use.
ComfyUI-Przewodo-Utils Features
Wan video tools
- WanVideoEnhanceAVideo: A video enhancement and post-processing node described as supporting quality improvement, artifact reduction, and consistency between frames. It is relevant when generated footage needs a finishing pass.
- WanVideoVaeDecode: Decodes video latents, the intermediate data a generation workflow uses before producing viewable frames. It is designed for video sequences and memory-conscious processing.
- WanModelTypeSelector: Selects between GGUF and Diffusion model types so a workflow can follow the appropriate path for the model in use. These are model types, not separate models supplied by this extension.
- WanVideoGenerationModeSelector: Selects a mode based on the images you want to guide a video with: start image, end image, start + end, or start + end + start. For example, choose start + end when you want to provide both the opening and closing visual references.
- WanGetMaxImageResolutionByAspectRatio: Calculates a suitable resolution for a Wan workflow while preserving the intended aspect ratio. This can help you prepare portrait or landscape inputs without unintentionally stretching them.
- WanVideoLoraStack: Organizes multiple LoRAs and their strengths for a video workflow. A LoRA is an additional model component used to influence a generation; adjusting its strength changes how strongly it is applied.
The extension describes multi-chunk support for creating sequences longer than a model’s usual limit, with attention to consistency and transitions between chunks. It also lists TeaCache, SageAttention, Block Swap, and TAESD Preview as performance or preview features. These may be useful when generation speed, memory use, or previewing matters, but the supplied documentation does not specify settings or guaranteed performance gains for them.
General workflow utilities
Choices and comparisons
- CompareNumbersToCombo: Compares two numbers using
==,!=,<,>,<=, or>=, then returns a string associated with the result. You could use it to choose text for another part of a workflow based on a numeric condition. - SwapAnyCondition: Uses a true/false condition to decide whether to swap two values. It is intended to work with different value types.
- SwapAnyComparison: Swaps values according to a numeric comparison.
- SwapImageComparison: Provides comparison-based swapping specifically for images.
- FloatIfElse: Chooses between two floating-point numbers based on a true/false condition. This can be useful when a setting needs one numeric value in one case and another value in the alternative case.
Images and sizing
- ImageScaleFactor: Calculates a factor for resizing an image while maintaining its aspect ratio. It helps when you want to change image size without distorting its shape.
- ImageSizer: Provides multiple image-sizing modes and creates an empty latent with an aspect ratio suited to the selected model type. An empty latent is the starting canvas, in an intermediate data format, for a generation workflow.
- BatchImagesFromPath: Loads multiple images from filesystem paths. Pattern matching and recursive directory scanning can help gather images for batch work instead of selecting them individually.
Lists and checks
- AppendToAnyList: Adds a single item to a list or combines lists and arrays. Use it when a workflow needs to collect several values before passing them onward.
- DebugLatentShapes: Displays latent dimensions and reports memory usage. This can help identify a size mismatch when connected nodes do not behave as expected.
- IsInputDisabled: Checks whether an input is disabled, allowing a workflow to respond to that state.
- HasInputValue: Checks whether an input contains a value rather than
None. This is useful before sending an optional input to another node.
The pack also describes tooltips, organized node categories, and error reporting intended to make its settings easier to understand and problems easier to diagnose.
ComfyUI-Przewodo-Utils Models
This extension is a node pack, not a collection of downloadable models. The supplied documentation identifies Wan2.1 as the main model family for its video tools. It also describes workflow support for GGUF and Diffusion model types through WanModelTypeSelector.
If you are choosing between those types, select the one that matches the model already used by your workflow. The provided documentation does not list individual Wan2.1 model files or prescribe a specific model for each creative task.
Troubleshooting ComfyUI-Przewodo-Utils
A node is missing from your workflow
- Look under both PrzewodoUtils and PrzewodoUtils/Wan in ComfyUI’s node menu.
- Check the exact node name against the feature list above.
- If the node still does not appear, consult the project repository for the extension’s current status and report the missing node through GitHub Issues.
A Wan workflow does not follow the expected model path
- Check which model type the workflow uses: GGUF or Diffusion.
- Confirm that
WanModelTypeSelectormatches that type. - Check the rest of the workflow for nodes or inputs that depend on the selected type. The original documentation identifies ComfyUI-GGUF as an enhancement for GGUF support.
An image or video has the wrong proportions
- Check the aspect ratio of your source image and the dimensions requested later in the workflow.
- Use
ImageScaleFactor,ImageSizer, orWanGetMaxImageResolutionByAspectRatiowhere appropriate to calculate dimensions before generation. - If the result still looks stretched or cropped, inspect any downstream resize or crop nodes; they may change the dimensions after these utility nodes run.
A workflow fails around latent data
- Connect
DebugLatentShapesat the point you want to inspect. - Compare the reported dimensions with what the next node expects.
- Recheck earlier sizing choices, especially the model type and aspect ratio used to create an empty latent.
Frequently asked questions
Does the extension include Wan2.1 models?
No models are listed as part of the node pack. The documentation describes utilities for workflows using Wan2.1.
Can I use the general utility nodes without making video?
Yes. The PrzewodoUtils category includes general image, list, comparison, and input-checking nodes.
Which video generation mode should I choose?
Choose according to the visual references you have: a start image, an end image, or both. The selector also offers a start + end + start mode. The supplied documentation does not describe its exact behavior beyond the mode name, so check the node’s tooltip when building your workflow.
Where can I report an error or ask about a node?
Use GitHub Issues for bugs and feature requests, or GitHub Discussions for questions and community conversation.
Learn More about ComfyUI-Przewodo-Utils
- Project repository: Explore the node pack and its documentation.
- Community discussions: Ask workflow questions and learn how others use the nodes.
- Bug reports and feature requests: Check known problems or describe an issue you encountered.
- ComfyUI: Learn more about the node-based application these utilities extend.
