OpenClaw skill routes plain English commands to ComfyUI batch jobs

OpenClaw's comfyui-skill-public routes natural-language commands to ComfyUI's API, turning requests like "upscale this folder to 4K" into queued batch jobs. The skill constructs workflows programmatically and tracks execution status across local, remote, and cloud deployments.

The tool targets users who want to script image processing without hand-editing JSON workflow files. OpenClaw parses English-language instructions, maps them to ComfyUI nodes, and submits the resulting workflow to the target instance's queue. That removes the need to manually wire sampler nodes, upscaler models, and file-save logic for routine batch operations.

What stands out

1. 01Natural-language interface — Commands are plain English rather than node-graph edits or API calls. The skill handles parsing and node selection internally.
2. 02Batch folder processing — Point the skill at a directory of images and it queues one job per file automatically, useful for upscaling or style-transfer runs on hundreds of outputs.
3. 03Multi-instance support — Works with local ComfyUI installs, remote servers over HTTP, and cloud-hosted endpoints from a single command. Users can route heavy jobs to a cloud GPU without changing the command syntax.
4. 04Automatic workflow construction — The skill assembles the node graph on demand rather than requiring pre-saved templates. That reduces setup friction for one-off tasks.
5. 05Job tracking — Monitors queue position and completion status for each submitted task, surfacing progress without requiring manual API polling.