comfyui-workflows

This repository is a collection of ready-made workflow files for ComfyUI, a node-based interface used to run AI image and video generation models. The README is in Chinese, but the content is straightforward. The workflows are saved as JSON files in an API format, meaning they are designed to be called programmatically or loaded into tools that support ComfyUI's API, rather than arranged for visual drag-and-drop use.

The collection covers a wide range of image and video tasks. For image work, there are workflows for generating product photos with AI backgrounds, swapping product backgrounds while keeping the subject intact, relighting portraits and objects, upscaling and repairing images to 2K, 4K, or 8K resolution, face swapping, clothing extraction and virtual try-on, outpainting an existing image beyond its edges, and generating anime or illustrated styles. For video, there are workflows built around the Wan 2.1 and Wan 2.2 video generation models, covering text-to-video, first-and-last-frame video, motion transfer, and animated character videos.

To use a workflow, you download or clone the repository, pick a JSON file from the workflows_api folder, replace the placeholder image or video paths with your own files, install any missing custom nodes using ComfyUI Manager, and then run it via the ComfyUI API or a compatible tool.

The files do not include model weights, API keys, or example images. Those must be obtained separately. The repository is described as a public sharing version of a personal workflow library and has no stated license.