Gradio Workflow Team Rebuilds AUTOMATIC1111 as a 73-Node Workflow Canvas

The Gradio Workflow team has released Workflow1111, a full reconstruction of AUTOMATIC1111's Stable Diffusion WebUI as a single graph-based workflow canvas. Built from 73 nodes across 11 media pipelines, it covers text-to-image, hi-resolution fix, image-to-image, ControlNet-style annotators, detection-to-inpaint masking, VLM image interrogation, background removal, PNG metadata storage, and image-to-video — all wired together using Gradio's four core operator types: Python functions, model calls via InferenceClient, Gradio Spaces, and Hub dataset rows.

Notable design choices include using FLUX.1-Kontext for hi-res upscaling and image editing, Qwen3-4B for LLM-assisted prompt expansion, and Qwen2.5-VL for image-to-prompt interrogation. Unlike ComfyUI, there are no custom node types — LLMs and diffusion models coexist as ordinary operators on the same canvas. The project runs via Hugging Face authentication, consuming the user's own inference quota, and can be duplicated as a starting point for custom pipeline development.