cardenluo/ComfyUI-Apt_Preset

What it solves

This plugin addresses the complexity of ComfyUI workflows, which often become dense and difficult to manage as they grow. It reduces visual clutter and operational errors by simplifying connections and standardizing the way users load models and control image generation.

How it works

The project reorganizes ComfyUI nodes into a structured "Loader — Controller — Sampler" framework:

• Universal Loaders: Provides advanced loaders (like Sum_load_adv) that support various model types (XL, SD, Flux, SD3.5, GGUF) and allow users to save model and sampling configurations as presets.
• Controller Stacks: Groups related control tools into centralized modules for specific tasks, such as general image control (ControlNet, Inpaint), video generation (Wan), animation (Animatediff), and multi-image reference (Kontext, QwenEdit).
• Simplified Samplers: Packages complex built-in sampler input ports into single-port replacements or combines basic sampling with special functions like secondary sampling repair and refine enlargement.
• Utility Nodes: Includes a suite of tools for processing data, images, and masks (including segmentation for faces, clothes, and bodies).

Who it–s for

ComfyUI users who want to organize their workflows more efficiently, reduce repetitive wiring, and use a more streamlined approach to managing multiple AI models and control tools.

Highlights

• Preset System: Save and reuse specific model and sampling combinations to avoid manual reconfiguration.
• Broad Model Support: Compatible with a wide range of architectures including Flux, SD3.5, and GGUF models.
• Modular Control: Specialized stacks for different modalities like video and animation.
• Integrated Processing: Built-in utility nodes for mask creation and image pre/post-processing.