What problem does it solve?
ComfyUI users often need to manually assemble large node graphs (models, samplers, encoders/decoders, and outputs) to create image/video/3D pipelines, which is time-consuming and error-prone.
This skill turns a natural-language request into a complete, importable ComfyUI Litegraph workflow JSON so you can run it immediately without manual node wiring.
Core Features & Use Cases
- Natural-language to workflow JSON: Generates valid ComfyUI workflow JSON (Litegraph format 0.4) that you can import and queue.
- Template-driven pipelines across modalities: Supports text-to-image, image-to-image, inpaint, upscale, audio generation, video generation (multiple families), and 3D workflows.
- Model-aware workflows: Automatically includes a top-level models array by extracting referenced model filenames, enabling ComfyUI to prompt downloads for missing models.
- LLM integration when needed: Can generate LLM-assisted pipelines using the
comfyui_LLM_party node package (provider/model selection guided).
Quick Start
Use the skill to generate a ComfyUI workflow for your task by writing a single request like: "Create a FLUX image-to-video workflow that turns my uploaded start image into a 5-second cinematic shot with camera motion and motion continuity."
