ComfyUI

Updated 2026-01-18: For NVIDIA GPU users, switching between different runtime environments can better leverage GPU performance and ensure compatibility with older models. For pre-50 series GPUs, use the default environment (torch 2.7.1) for stability. For 50 series and newer GPUs, torch 2.8.0 or even torch 2.9.1 is recommended. Note: switching environments triggers dependency checks, which can be time-consuming. Sticking to one environment ensures faster operation. "How to choose a virtual environment for ComfyUI?"

ComfyUI is a node-based graphical user interface for stable diffusion, mainly used to build highly flexible image generation workflows, and is suitable for various user groups such as digital artists, AI researchers, and content creators.

• Core Features:
  • Lightweight and Fast-running: Compared to some traditional image generation tools, ComfyUI has relatively low resource consumption, can run smoothly on devices with lower configurations, and the speed of generating images is also relatively fast, which greatly improves the creative efficiency.
  • Highly Flexible and Configurable: Users can build various complex image generation workflows by linking different nodes. Each node represents a specific function or operation, such as loading checkpoint models, inputting prompts, specifying samplers, etc. Users can freely combine and adjust these nodes according to their own needs to achieve personalized image generation effects.
  • Transparent and Intuitive Workflow: ComfyUI presents the image generation workflow in a visual way of nodes and connections. Users can clearly see the flow of data and the processing results of each node. This helps users better understand and master the principle and process of image generation, and also facilitates users to debug and optimize the workflow.
  • Easy to Share and Spread: The workflows of ComfyUI can be saved as files, and each file is a repeatable workflow. Users can easily share the workflows they have created with other users, and they can also download and use the workflows shared by others, which promotes the exchange and sharing of knowledge and experience.
  • Reproducibility: By saving and calling workflows, users can ensure that the generated images are consistent each time, which greatly improves the stability and predictability of work.
  • Rich Plugin Support: It supports the installation and management of various plugins. Users can add new functions or optimize existing functions according to their needs, further improving work efficiency.
• System Support: Supports Windows, macOS, and Linux platforms.
• Installation and Use: You can use LM Downloader to quick install, or downloade integration package provided by the official website and follow the instructions in the official documentation to download, decompress, and configure it. After the installation is completed, open the starter file of ComfyUI to start the program. After starting, a web-based UI interface will be opened in the browser. The interface mainly consists of nodes, connections, and control areas. Users can choose existing workflows or create new ones. After configuring the workflow parameters, click the "Queue Prompt" button or other relevant running buttons on the interface, and ComfyUI will start to generate images according to the set workflow and parameters, and display the generation progress and results on the interface.