The current workflow connections are becoming increasingly dense, making them difficult to understand and operate, and there are issues with poor consistency. As the number of saved workflows increases, frequent errors occur during actual use, and the operation methods need to be relearned. To address this, a plugin has been developed to simplify workflows, clarify control concepts, and design nodes around the "Loader ---- Controller ----- Sampler" framework.

<a href="./README.md">English</a> >> <a href="./README_ZH.md">中文版</a>

I. Update Record

2025.1.14 Optimize sum_load to support caching and cleaning, update AI_GLM to support the new GLM 4.7 model, and add judgment nodes to the process group

2025.01.17 sum_load and sum_edit are compatible with the latent size of Flux2_Klein

2025.1.25 Add an image batch loading manager：IO_LoadImgList

2025.1.31 update IO_LoadImgList Supports bridging, with more powerful operations

2026.3.13 add LTX2.3 video_merge: AD_video_merge

2026.3.15 Add visualization batch loading nodes ：IO_LoadImgBatch、IO_LoadTextBatch、IO_LoadShotBatch

2026.3.22 Add image editing and visualization nodes

<img width="2028" height="868" alt="image" src="https://github.com/user-attachments/assets/bd2f2bb4-b82d-4297-bff8-cb40f46bbff2" /> 2026.3.29 Add Automatic Video Segmentation Node ："AD_VideoSeg": AD_VideoSeg,

2026.4.27 Updated the prompt visualization tool “text_interPrompt“” to support 4 modes

II. Usage Guide

1. Universal Loader Sum_load_adv, Supports GGUF Models

Model combinations for various workflows are the same as official workflows.

① XL, SD mode: Load model checkpoint or Unet or over model.

② wan2.1, wan2.2, QwenEdit mode: Load model Unet + clip1

③ Flux\Kontext mode: Load model Unet + clip1 + clip2 (Note: In order, not clip1+clip3)

④ SD3.5 mode: Load model Unet+clip1 +clip2+clip3

⑤ Hi-dream mode: Load model Unet+clip1 +clip2+clip3+clip4

Override mode over_model and over_clip

① When using over_model, the internal corresponding model or Unet of the loader becomes invalid and will directly output model

② When using over_clip, all internal corresponding clip1,2,3,4 of the loader become invalid and will directly output clip

Preset saving, which can save loaded models and sampling methods uniformly

① At least one arbitrary preset must be selected, otherwise an error will occur

② After selecting any preset, all parameters can be modified and take effect arbitrarily.

③ Newly set parameters can be saved as new presets, but ComfyUI needs to be restarted to select and use them

2. Controller Stack: Functional modules with associated control tools centralized together

① General Control image Stack: Universal control for SD, XL, Flux, etc., Ipa style, redux transfer, Union_controlnet, controlnet_adv, inpaint redrawing

② General Control wan Stack: All wan video generation nodes officially supported

③ General Control AD Stack: Animatediff animation generation control, prompt scheduling, CN scheduling, IPA scheduling, etc.

④ General Control Kontext Stack: Multi-image reference, partitioned generation, redux transfer, union_controlnet

⑤ General Control QwenEdit Stack: Multi-image reference, union_controlnet

3. Sampler: Rich sampling methods to eliminate repetitive connections

① Basic sampler: Packages ComfyUI's built-in sampler input ports into a single port replacement, with functionality completely consistent with the official version

② Special function sampler: A combination of basic sampler + special functions.

As in the example below, operations like secondary sampling repair and refine enlargement can be achieved in one step:

4. Utility Nodes: Image, mask, data processing tools

Data types: Data conversion, data creation，data operation

Image processing: A powerful image processing combination: pre-processing, intermediate generation, and post-recovery

Mask processing: Creation, conversion, and transformation of masks

III. Installation

Clone the repository to the custom_nodes directory and install dependencies


#1. Git download
git clone https://github.com/cardenluo/ComfyUI-Apt_Preset.git

#2. Install dependencies
Double-click install.bat to install dependencies

Note:

1、To use the controlNet schedule control feature, please first install ComfyUI-Advanced-ControlNet

2、To load GGUF models, please first install ComfyUI-GGUF

3、To use the load_Nanchaku node, please first install and make sure it works normal ComfyUI-nunchaku.

4、ollama model: download and put at "...comfyui\models\ollama" Download model.

5、<a id="AAA">Some nodes will use the resource expansion pack. Please download it : Apt_file. or download it here: Mask_FaceSegment 、Mask_ClothesSegment、 Mask_BodySegment Then Place the entire folder into comfyUI/models</a>

├── ComfyUI/models/Apt_File
|     ├──body_segment
├── ComfyUI/models/Apt_File
|     ├──segformer_clothes
├── ComfyUI/models/Apt_File
|     ├──segformer_face

Disclaimer

This open-source project and its contents are provided "AS IS" without any express or implied warranties, including but not limited to warranties of merchantability, fitness for a particular purpose, and non-infringement. In no event shall the authors or copyright holders be liable for any claim, damages, or other liability, whether in an action of contract, tort, or otherwise, arising from, out of, or in connection with the software or the use or other dealings in the software.

Users are responsible for ensuring compliance with all applicable laws and regulations in their respective jurisdictions when using this software or publishing content generated by it. The authors and copyright holders are not responsible for any violations of laws or regulations by users in their respective locations.

Follow Me, Share Methods for Easily Building Workflows

Bilibili: https://space.bilibili.com/2008798642?spm_id_from=333.33.0.0