comfyui-local_skill

Overview

This skill connects to a locally running ComfyUI instance to generate high-quality images using your own hardware and custom workflows. It prioritizes privacy and control by sending prompts and workflow JSON directly to your ComfyUI server. Requires a running ComfyUI server on the local network and Dev/API mode enabled. Outputs are saved locally for easy backup or sync.

How this skill works

The skill uses a Python helper script to perform the HTTP/WebSocket handshake with the ComfyUI API, enqueue a prompt, and download the finished image. By default it loads a high-quality SDXL workflow but can inject prompts and random seeds into any JSON workflow placed in the workflows folder. Images are written to a local image-gens folder and can be configured to sync with your document directories. Server address and API access are controlled through environment variables and ComfyUI Dev mode.

When to use it

• You want private image generation on your own GPU rather than a cloud service.
• You need reproducible results using custom ComfyUI workflows (JSON).
• You want to run high-quality SDXL workflows locally for production assets.
• You need images automatically saved and backed up to local folders.
• You want to integrate ComfyUI into scripts or automation on your LAN.

Best practices

• Set COMFYUI_SERVER_ADDRESS to the exact IP and port of the machine running ComfyUI.
• Enable ComfyUI Dev mode to allow API interactions before running the script.
• Store and version your workflow JSON files in the workflows/ folder for reproducibility.
• Use the default SDXL workflow for best-quality outputs, customize only when needed to avoid breaking node structure.
• Keep output folder (image-gens/) writable and configure a separate sync/backup job for long-term storage.

Example use cases

• Quickly generate production-ready SDXL images on a local workstation for private projects.
• Run a scheduled script that produces daily image variations and saves them to a synced folder.
• Test and iterate custom ComfyUI workflows by dropping JSON files into workflows/ and invoking the script with a prompt.
• Integrate local image generation into a content pipeline where assets must never leave the local network.
• Generate images with a reproducible seed injected automatically for batch comparisons.