Command-line interface for ComfyUI. Queue workflows, monitor jobs, download results — all from your terminal.
Built for headless servers, SSH sessions, and automation pipelines. No browser required.
ComfyUI's web UI is great for building workflows. But once you have a workflow that works, you want to:
- Queue it from a script or cron job
- Monitor GPU jobs from another machine via SSH
- Download results without opening a browser
- Batch-process multiple prompts overnight
- Integrate with other tools and pipelines
This CLI does all of that.
pip install comfyui-cliOr run directly:
python -m comfyui_cli --help# Check server status
comfyui-cli status --host your-server:8188
# Queue a workflow
comfyui-cli queue workflow.json --host your-server:8188
# Queue with parameter overrides
comfyui-cli queue workflow.json --set "KSampler.seed=42" --set "KSampler.steps=30"
# Watch running jobs
comfyui-cli watch --host your-server:8188
# Download latest outputs
comfyui-cli download --host your-server:8188 --output ./results/
# List available models
comfyui-cli models --host your-server:8188- Queue workflows from JSON files with parameter overrides
- Monitor running/pending jobs with progress bars
- Download generated images and videos
- List available models, LoRAs, and checkpoints
- Batch mode for overnight processing
- Remote server support (any ComfyUI instance on your network)
- Scriptable output for piping to other tools
Save your server details so you don't need --host every time:
comfyui-cli config set host your-server:8188
comfyui-cli config set output_dir ~/comfyui-output/MIT