Comfy MCP Pipeline

What is Comfy MCP Pipeline?

Comfy MCP Pipeline is a pipeline wrapper for comfy-mcp-server that integrates with Open WebUI. It lets you run image generation using ComfyUI workflows directly from an Open WebUI chat interface.

How to use Comfy MCP Pipeline?

Upload the comfy-mcp-pipeline.py script to your Open WebUI Pipelines server, select the “comfy-mcp-pipeline (pipe)” valve, and configure the required valves: Comfy Url, Comfy Url External, Comfy Workflow Json File, Prompt Node Id, and Output Node Id. Then start a new chat, select “Comfy MCP Pipeline” as the model, enter a text prompt, and send to generate an image.

Key features of Comfy MCP Pipeline

• Wraps comfy-mcp-server for Open WebUI Pipelines
• Uses a JSON export of a ComfyUI workflow (API format)
• Configurable valves for server URL, workflow file, and node IDs
• Prompt-driven image generation from chat

Use cases of Comfy MCP Pipeline

• Generate images via Open WebUI using custom ComfyUI workflows
• Automate image creation by triggering pipeline with text prompts
• Deploy stable diffusion workflows in a chat-based environment

FAQ from Comfy MCP Pipeline

What prerequisites are required?

You need Open WebUI, Open WebUI Pipelines, ComfyUI, an updated requirements.txt on the pipelines server, and a JSON export of a ComfyUI workflow.

How do I export a ComfyUI workflow for the pipeline?

In ComfyUI, select your workflow, go to the top menu → Workflow → Export (API), provide a filename, and confirm. Upload the resulting JSON file to the pipeline server.

What valves must be configured?

The required valves are: Comfy Url (your ComfyUI server), Comfy Url External (use same if internal), Comfy Workflow Json File (path on server), Prompt Node Id (text prompt node), and Output Node Id (generated image node).