zho-zho-zho/comfyui-workflows-zho

This is a personal collection of workflow configuration files for ComfyUI, a node-based tool used to run AI image and video generation models locally. ComfyUI lets users connect processing steps visually, and workflows are JSON files that encode those connections so others can reproduce the exact same setup without rebuilding it from scratch. The collection contains 54 workflows across 22 categories, assembled by a creator who publishes tutorials on the Chinese video platform Bilibili. The categories cover a wide range of image and video generation tasks. Image generation workflows include Stable Cascade, FLUX.1 Dev and Schnell, Stable Diffusion 3, and various image-to-image and inpainting setups. Identity-preserving generation workflows use models like InstantID and PhotoMaker to generate new images that keep a specific person's face consistent across outputs. Portrait animation workflows include LivePortrait for animals and a Hunyuan Video setup for generating short video clips. There are also workflows for converting sketches into 3D objects, combining large language models with image generation, applying differential diffusion for region-specific editing, and running object detection and segmentation through YoloWorld and EfficientSAM. Each workflow entry in the README shows a preview image or video of the output, along with a link to download the corresponding JSON file. The files can be loaded directly into a local ComfyUI installation. Some entries also link to Google Colab notebooks for running the workflow without a local GPU setup. The repository does not contain any Python code or training scripts. It is purely a library of ready-to-use workflow files for people who already have ComfyUI running and want to try specific generation techniques without having to wire up the node graph themselves.