A simple base front-end for text-to-image workflow in ComfyUI. Meant to serve as a base to be modified for future complex workflows
Inspired by https://www.youtube.com/watch?v=anYHG37fUg4 , but some of his endpoints didn't quite work the same for me, so I modified them by following some example usage from ComfyUI Github discussions: Comfy-Org/ComfyUI#1037
Building a front-end UI for ComfyUI workflows using their API. The front-end will be able to take in text input and display the generated image output.
This is a simple UI meant to be a base for future more complex workflows.
Simply git clone this repository in your ComfyUI/custom_nodes directory.
This does not require any external python packages
The workflow is embedded in the file web/js/app.js in the workflow variable.
The initial plan was to have it load from an external json file but for some reason, the changes made to json were not changing during server restarts, so for sake of consistency, the workflow is embedded in the file.
Feel free to move to a .json file and pull it via fetch if the workflow is too large.
To modify the workflow, simply find the appropriate settings in the workflow variable and modify them.
Some notable settings:
sampler_name,scheduler,steps,cfg(KSampler)ckpt_name(CheckpointLoaderSimple) : This needs to be the exact filename of the checkpoint in ComfyUi/models/checkpointswidth,height(EmptyLatentImage)
- Start the ComfyUI server with
python main.py --listen 0.0.0.0 --enable-cors-header '*' - Find your computer's IP address and navigate to
http://<your-ip>:8188/testuito start using it
- The workflow is embedded in the
app.jsfile. This is not ideal and should be moved to a separate file and fetched viafetchfor better maintainability. - It's using Tailwind CDN. Didn't want to go through the hassle of setting up a build system for this simple project. This should be changed to a build system for production.

