About

Comfy Modder allows you to replace inputs in your ComfyUI workflows.

• Free
• Open source
• Runs in your browser
• Interacts with your local ComfyUI process via websocket

Example use case

Imagine you want to generate images of a landscape at different times of day using both prompt variation and various LORAs.

| prompt    | LORAs                           |
| --------- | ------------------------------- |
| sunrise   | Darkness wt: 0; Sunshafts wt: 1 |
| midnight  | Darkness wt: 1; Sunshafts wt: 0 |

Using an XYZ Plot that iterates over each input would generate extra images that you don't want, such as sunrise + Darkness wt: 1

Comfy Modder addresses this use case by enabling generation of only the desired inputs.

Usage

1. Install ComfyUI
2. Add --enable-cors-header * to the commandline options of run_nvidia_gpu.bat
3. Start ComfyUI by running run_nvidia_gpu.bat
4. Create a workflow that generates an image
5. Enable devtools in the ComfyUI settings menu
6. Save the workflow in API format
7. Copy and paste the API formatted workflow into comfy modder
8. Configure replacements that you would like to change in your workflow
9. Click generate

Tips 💡

• Images are not saved to disk automatically unless your workflow does so. Make sure to download generations before refreshing the page.
• Dropdown selector fields filter options based on the text entered - if you aren't seeing an option you expect, delete the current text in the field.