Rebuilding AUTOMATIC1111 with Gradio Workflow
| Source: Hugging Face Blog
Tags: Hugging Face, Gradio, AUTOMATIC1111, Stable Diffusion, FLUX, image generation, Workflow
Hugging Face rebuilt AUTOMATIC1111's full stable diffusion UI as Workflow1111 — a single Gradio Workflow canvas of 73 nodes spanning 11 pipelines including text-to-image, FLUX.1-Kontext hi-res fix, img2img, ControlNet annotators, inpainting, and image-to-video — runnable by any HF account holder against their own quota.
Details
Workflow1111 demonstrates what Hugging Face's Gradio Workflow graph system can do at the complex end: 73 nodes, 11 media pipelines, replicating most of AUTOMATIC1111's feature set without writing orchestration code. The graph uses four operator types: fn (Python function), model (InferenceClient call), space (another Gradio Space), and dataset (Hub dataset row). Pipelines covered include: text-to-image with style presets and PNG metadata embedding; hi-res fix via FLUX.1-Kontext with a 'enhance fine detail, keep composition identical' instruction; image-to-image via the same Kontext node; LLM-assisted prompt expansion via Qwen3-4B (output capped at 40 tags); detection-based inpaint masking; ControlNet-style annotators; background removal; image-to-video; and PNG Info reading. The practical significance: tasks that previously required custom Python orchestration scripts or fluency with ComfyUI workflow JSON can now be built visually on a shared canvas. The space is publicly accessible and can be duplicated and rewired by any user. Compute runs against the authenticated user's own HF quota. The Gradio Workflow API is documented and open — this release functions partly as a demonstration of what the graph system can handle, and partly as a starting template for practitioners who want to fork and extend it.