yanokusnir-ai/one-node-flux-2-klein

What it solves

It eliminates the need to build complex, "spaghetti-wire" node graphs in ComfyUI when using FLUX.2 [klein]. By wrapping an entire professional workflow into a single UI widget, it allows users to access advanced image generation and editing features without manual node routing.

How it works

The project is a ComfyUI custom node that provides a unified interface for the FLUX.2 [klein] model. It integrates multiple operational modes—including text-to-image, image-to-image, and specialized editing tools—into one node. It supports external model loaders (including GGUF) and integrates with other custom nodes for specific tasks like inpainting and pose estimation.

Who it’s for

Digital artists and AI image creators who use ComfyUI but want a more streamlined, "all-in-one" experience for FLUX.2 [klein] without the overhead of managing large graphs.

Highlights

• Seven Integrated Modes: Supports T2I, I2I, EDIT, PAINT (Sketch, Inpaint, Outpaint), FACESWAP, POSE, and UPSCALE.
• Advanced Painting Tools: Includes a full canvas with layers, brushes, and shapes, as well as reference-guided inpainting.
• Pose Control: Ability to copy a pose from one image onto a character from another using DWPose and RefControl LoRAs.
• High-Quality Upscaling: Integrated SeedVR2 upscaler for enlarging images up to 8x.
• Flexible Model Loading: Supports both 4B and 9B FLUX.2 [klein] variants and allows for external model/CLIP/VAE inputs.