Comfy CLI for OpenClaw to run StableDiffusion workflows
A developer built a CLI that bridges OpenClaw to ComfyUI for StableDiffusion image generation. The agent describes what it wants in plain language, and the CLI handles the complex workflow JSON.
What the CLI Does
- Workflow management: list, select, switch between saved ComfyUI workflows
- Parameter editing: change prompt, negative prompt, steps, seed, checkpoint without raw JSON
- Execution: trigger runs and monitor progress
- Output: generated images saved and accessible to the agent
How It Works
The CLI abstracts ComfyUI's complex JSON format. Agent says "generate an image of X" and the CLI translates to ComfyUI parameters. Open-sourced on r/StableDiffusion.
Results: Working OpenClaw-to-ComfyUI bridge. No raw JSON needed. Open-source.