Dear Photo Lover,
You found a prompt that makes a picture you love. Now what?
Most of us save it, run it again, and move on. WorkflowPad does something more fun: it takes that picture, or its prompt, as the first frame of a story, and grows it. The same character, the same place, a new pose and a new feeling on every shot, until one lucky picture has become a little illustrated tale.
Pick a mood, paste your favourite prompt, and let your own ComfyUI paint the rest while you sit back and enjoy the show.
This is the idea in one picture: the prompt came from https://civarchive.com/images/142275949, and WorkflowPad built a story around it. The line on the picture is the narration the program writes onto it. Each shot of a story gets its own line.
Want to try it? WorkflowPad is free and open source. Download the program (Windows, portable): https://github.com/sudja-tongtip/WorkflowPad/releases
Turn a story into a picture book, with your own ComfyUI
WorkflowPad is a free, open-source set of five desktop programs for Windows that sit on top of ComfyUI. You write a story, or a single idea, and it plans the shots, writes the image prompts, and sends them to your own workflows. These are the starter workflows that make it run.
Story Generator: the story on the left, the storyboard in the middle, rendering on the right.
Why it's fun
Story to pictures. Write a few lines, press a button, and watch a story become a Story Bible, a Beat Sheet, a Shot List and finally pictures, one shot at a time. Review and edit any step before you go on.
Every shot has its own mood. New pose, new expression, new feeling on each shot, so the set reads like a tale and not like one picture repeated.
Narration on the picture. Each shot can carry a line of text, like the ones you see here.
Manga mode. Whole pages with panels, so you get a finished page, not just loose pictures.
Bring your own workflow. Edit your ComfyUI workflows however you like, drop them in, and have fun. WorkflowPad finds the right nodes for you, so any model you love can power the show.
Example prompt adapted from https://civarchive.com/images/141319690
Fix it, don't redo it
Pictures not right? Don't start over. WorkflowPad is five small desktop programs, and each one fixes a different kind of problem:
• Image Editor: open a folder of pictures and work on them in one window. Create a new picture, edit one with a sentence, repaint just a masked part, extend the edges of the frame, restore a face, upscale, or turn a picture back into a prompt. Zoom, pan and multi-select like in a file browser.
• Prompt Lab: a notebook for your prompts. Ask an LLM (a local model through ComfyUI, or Grok) to rewrite them, then run the result straight through an image workflow.
• Batch Editor: got a whole folder? Pick one workflow and one prompt, run it across all the pictures, or select just the ones you want.
• Gallery Generator: want endless variations? Give it a list of prompts with {a|b|c} wildcards and it keeps filling a gallery of thumbnails while you watch.
• Story Generator: the one at the top of this page. It turns a story into a Story Bible, a Beat Sheet, a Shot List and finally the pictures.
What you get in this download
Seven example ComfyUI workflows:
Text to image (Krea2)
Image edit (Flux2, no mask)
Masked edit / inpaint (Flux2)
Extend / outpaint (Flux Fill)
Face restore and upscale
Image to prompt (JoyCaption)
Local LLM chat (Gemma) for story writing
How to use
Drag a .json onto the ComfyUI canvas.
Pick the models you own in each loader, then run once to confirm it works.
In ComfyUI choose Workflow > Export (API) and save the file.
Put the file in the matching folder of WorkflowPad's comfy_apis folder (image_create, image_edit, enhance, caption, llm).
Open WorkflowPad. Your workflows show up in the dropdowns.
Results you can expect
You describe a scene once and get a run of shots built around the same character and place, with a mood that changes from shot to shot instead of one pose repeated. Edits stay local, so a fix to one picture does not disturb the rest of the set.
Good to know
Needs a recent ComfyUI (the examples use Subgraphs). If it asks for missing nodes, install them from ComfyUI-Manager: ComfyUI-GGUF, ComfyUI-Inpaint-CropAndStitch, rgthree-comfy, the face-restore nodes, the JoyCaption GGUF nodes and the LLMSessionChatNode.
Models are not included. Check each model's own license.
The outpaint example contains a LoRA node that does not suit Flux Fill; bypass it if ComfyUI warns.
Beta software: tell me what breaks.
Free and open source (AGPL-3.0).
Download the program (Windows, portable): https://github.com/sudja-tongtip/WorkflowPad/releases
Source code, manual and workflow notes: https://github.com/sudja-tongtip/WorkflowPad