demobook

ComfyUI: Generate an image-to-image workflow from a prompt

Demo summary

The demonstrator uploads an image and asks the agent to create a workflow to remove an object; the agent generates a Flux-based workflow which is then successfully executed.

Step-by-step

  1. Upload the image containing the object you want to remove
  2. Write a prompt specifying the removal of the target object
  3. Generate the workflow via the Copilot agent
  4. Open the generated Flux Dev workflow to review the node composition
  5. Execute the workflow without changing additional settings to see the result

Options

  • Use Qwen Image instead of the default Flux Context workflow

Watch out for

  • Invoking the 'generate workflow' function can be difficult as the agent may default to recommending existing workflows instead

Tips

  • Review the basic composition of the nodes before running to understand the workflow logic

Highlights

The overall effect is quite good

All demos from “Meet Your comfyUI AI Agent (RunningHub Edition)|Auto-Generate Workflows & Find Nodes Instantly!

  1. 1:590:54Search and add nodes with ComfyUI CopilotThe user demonstrates using the ComfyUI Copilot agent to find a specific node for calculating audio duration and adding it directly to the canvas.ComfyUIAI Image Generator
  2. 5:250:51Retrieve community workflows in RunningHubThe presenter uses the AI agent to search for 'VACE' and 'Infinite Talk' workflows from the RunningHub database and imports a selected workflow into the workspace.RunningHubAI Image Generator
  3. 7:330:55Generate an image-to-image workflow from a promptCurrentThe demonstrator uploads an image and asks the agent to create a workflow to remove an object; the agent generates a Flux-based workflow which is then successfully executed.ComfyUIAI Object Remover/Eraser
  4. Watch “Meet Your comfyUI AI Agent (RunningHub Edition)|Auto-Generate Workflows & Find Nodes Instantly!” →

AI Object Remover/Eraser

  1. 7:330:55Generate an image-to-image workflow from a promptCurrentThe demonstrator uploads an image and asks the agent to create a workflow to remove an object; the agent generates a Flux-based workflow which is then successfully executed.Veteran AI