ComfyUI-MakkiTools Introduction
ComfyUI-MakkiTools is a collection of custom nodes for ComfyUI. The author created these nodes to handle small, practical tasks that were needed in workflows but were not readily available among the nodes the author had found.
The collection covers image handling, translation, timing and information display, random numbers, integer statistics, LoRA loading, and Python package installation. For example, you can select one image from a sequence, join image batches, inspect the type of data passing through a workflow, or measure how long a process takes.
The nodes are separate tools, not a single image-generation pipeline. You can add the ones that suit your workflow without using the rest.
How ComfyUI-MakkiTools Works
A ComfyUI workflow passes information between connected nodes. Think of each MakkiTools node as a small workstation: it receives something, performs one task, and passes the result along.
- An image node might select, resize, separate, merge, or stitch images before another node uses them.
- An information node helps you inspect a workflow without changing its creative goal. For example,
show_typetells you what kind of data it receives, whiletimermeasures running time. - A translation node helps you work across languages. One translation option uses an M2M100 model; another integrates a multilingual translation tool.
- A LoRA node helps load multiple LoRA models as part of a workflow.
An image sequence is a set of images you can select from. An image batch is a group of images passed through a workflow together. MakkiTools includes nodes for both situations, so you can choose one image when you need a specific result or combine batches when you need to work with a larger group.
ComfyUI-MakkiTools Features
Image Manipulation
| Node | What it does | When you might use it |
|---|---|---|
GetImageNthCount | Retrieves the Nth image from an image sequence. | Choose a particular result from several images. |
ImageChannelSeparate | Separates a specified channel from an image. | Work with one channel independently. |
MergeImageChannels | Merges different image channels. | Bring separately handled channels back together. |
ImageCountConcatenate | Concatenates multiple image batches. | Pass images from separate batches onward as one combined batch. |
AnyImageStitch | Stitches any number of images using a specified dimension and reference size. | Assemble several images into a larger composition. The dimension and reference-size choices determine how the images are arranged and sized in that composition. |
Image_Resize | Resizes an image. | Prepare an image for a later step that calls for a different size. |
Prism_Mirage | Performs specific processing on two images. | Explore a two-image effect. The source description does not specify the processing method, so inspect the node’s inputs and preview its result before relying on it in a finished workflow. |
ImageChannelSeparate and MergeImageChannels are useful as a pair when you want to handle parts of an image separately. AnyImageStitch is the current stitching option listed by the author; ImageWidthStitch and ImageHeigthStitch, which stitch horizontally and vertically respectively, are deprecated. Keep that in mind when revisiting older workflows that use them.
Video Processing
AutoLoop_create_pseudo_loop_video converts a non-looping video into a pseudo-looping video. It is deprecated, so treat it as a legacy node rather than the first choice for a new workflow.
Information and Workflow Inspection
Environment_INFOdisplays system environment information. It can help you understand the environment in which your workflow is running.show_typedisplays the type of input data. Use it when you are unsure what a connection is carrying.timerdisplays the running time of a process. It can help you compare how long different parts of a workflow take.
Translation
translatorsintegrates a multilingual translation tool.translator_m2m100performs multilingual translation using an M2M100 model.
These nodes can help when text in your workflow needs to move between languages. The author notes that translation functions require a working network connection.
Numbers, Statistics, and LoRAs
random_anygenerates random integers and floating-point numbers. Use it when a workflow calls for a varying numeric value rather than the same value every run.int_calculate_statisticscalculates statistics for integers. It is intended for workflows that need a summary of integer values.BatchLoraLoaderloads LoRA models in batches. It is useful when your workflow needs to handle multiple LoRAs.
Package Installation
UniversalInstaller installs specified Python packages. This is a workflow maintenance tool rather than an image-editing effect. After installing a package, fully restart ComfyUI before expecting the change to take effect.
ComfyUI-MakkiTools Models
The only specific translation model named in the provided documentation is M2M100, used by translator_m2m100 for multilingual translation. The documentation does not list M2M100 variants or describe differences in their output or performance.
BatchLoraLoader works with LoRA models, but the author does not list particular LoRAs as part of MakkiTools. Your choice of LoRA is separate from the choice to use this loader. The other listed nodes are described as tools or functions, not as distinct image-generation models.
What's New with ComfyUI-MakkiTools
August 29, 2025: Node Index Renaming
The author renamed all node indices to reduce conflicts between MakkiTools nodes and other nodes, including future additions.
This matters most for existing workflows: after updating, a workflow may report that a MakkiTools node cannot be found. The author’s instruction is to pull that node again. Check older workflows for missing nodes and reconnect them as needed before continuing your work. No other dated changes are described in the provided documentation.
Troubleshooting ComfyUI-MakkiTools
A Node Is Missing After an Update
- Check whether the workflow uses a MakkiTools node.
- If it does, note that all node indices were renamed on August 29, 2025.
- Follow the author’s instruction to pull the missing node again.
- Check the restored node’s connections before running the workflow.
Translation Is Not Working
- Confirm that your network connection is working. The author specifically notes that the translation functions need a normal network connection.
- Check that the translation node is receiving the text you intended to translate.
- Try the translation step again once connectivity is restored.
An Installed Python Package Is Not Available
- Let
UniversalInstallerfinish installing the package. - Fully close and restart ComfyUI.
- Try the workflow again. The author notes that package changes require a full restart to take effect.
An Older Image or Video Node Causes Problems
Check whether the workflow uses ImageWidthStitch, ImageHeigthStitch, or AutoLoop_create_pseudo_loop_video. All three are marked deprecated. For image stitching, consider AnyImageStitch, which supports any number of images and offers dimension and reference-size settings. The provided documentation does not name a replacement for the deprecated video node.
Frequently Asked Questions
Do I need every MakkiTools node in one workflow?
No. The nodes perform separate tasks. Use only those that help with the workflow you are building.
Which node should I use to combine several images?
Use AnyImageStitch to assemble images into a stitched composition. Use ImageCountConcatenate when you want to combine image batches instead.
Can I use MakkiTools to find out why a workflow connection is confusing?
show_type can display the type of data it receives. Environment_INFO displays environment information, and timer can show how long a process takes.
Does MakkiTools include a list of recommended LoRAs?
No specific LoRA models are listed in the provided documentation. BatchLoraLoader is a tool for loading LoRAs in batches.
Learn More about ComfyUI-MakkiTools
- ComfyUI-MakkiTools repository: The extension’s source and project information.
- Chinese README: The author’s Chinese-language introduction.
- ComfyUI repository: Background on the application in which these nodes run.
