comfyui-missing-tool

comfyui-missed-tool

A few tools for ComfyUI, perhaps it's exactly what you need.

Installation

To install comfyui-missed-tool, clone the repository and install the dependencies:

git clone https://github.com/ifmylove2011/comfyui-missed-tool.git
cd comfyui-missed-tool
pip install -r requirements.txt

Table of contents

• Image Queue Loader
• Load Image alpha
• TrimBG
• TxtSave
• TextSplit
• ScaleMultiPixels
• Lora Merge
• ColorImageFillRm
• Face Similarity Check

ImageQueueLoader

Load Images in Queue, it is possible to obtain metadata information, such as seed, from the generated images in a queue. We can retrieve seeds sequentially during sequential execution.

LoadImagealpha

Load Image with alpha, it works when transparent images exits. Built-in load-image extension output is "RGB" default, not "RGBA". Of course, we can also use the built-in load image of Comfyui, but it needs to be used in conjunction with a mask.

TrimBG

We can also batch read images and process them. Better with Load Image List From Dir, list, not batch. And, Better to use with ComfyUI_BiRefNet_ll

TxtSave

Save txt file with input file path and name if "src_mode" enable.

TextSplitToList

Splitting a text or string into a list, you can use regular expressions to cut any string and output the list and the result after removing the delimiter.

ScaleMultiPixels

Scale image to multiple pixels. A input picture pixels is 533 * 433, if base_pixel set 100, output will be 500 * 400.

LoraMerge

Load Lora, Merge Load, and Save Lora. Combination is relatively more flexible, correspondingly, there are more nodes, totaling three!

ColorImageFillRm

Simple color replacement and removal of color pixels are useful in certain scenes where mask images are overlaid with layers.

FaceSimilarityCheck

Compare an input face against the images in a reference directory. The nodes return similarity scores, the best-matching filename, a readable Top-K report, and JSON results for use in downstream workflows. Reference embeddings are cached in cache_dir to speed up later runs.

• Face Similarity Check uses face_recognition embeddings and cosine similarity.
• Face Similarity Check (Insight) uses the InsightFace buffalo_l model.
• Face Similarity Check (Hybrid) ranks with face_recognition similarity and applies an ArcFace penalty to obvious mismatches. Configure arcface_penalty_threshold and penalty_factor when you need stricter rejection.
• Face Identity Similarity (antelopev2) builds a 512-dimensional identity centroid from reference_dir, can remove reference outliers, and returns similarity, passed, best_match, an aligned face crop, and a detailed report. This node requires the CUDA-enabled ONNX Runtime provider.

Set source_dir (or reference_dir) to a directory containing reference images. PNG, JPG, JPEG, WebP, and, where supported, BMP files are accepted. A clear, front-facing reference set generally produces the most stable results.

Contributing

If you want to contribute to comfyui-missed-tool, please fork the repository and create a pull request with your changes.

License

comfyui-missed-tool is licensed under the MIT License. See the LICENSE file for more details.

Contact

For any questions or feedback, please open an issue on GitHub or contact the repository owner.

Enjoy.