You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm a bit unclear about what the expected UX is for this application.
There isn't any way to actually access runpod deployed UI via a browser right?
Is the methodology described in "How to get the workflow from ComfyUI?" the only current supported method?
When is says "Open ComfyUI in the browser", it's implied that you are running a separate ComfyUI server somewhere(locally perhaps).
Can I run the UI locally, and have it render images on the runpod worker automatically when I press the queue button?
Or do I need to copy paste the workflow json each time I change something?
There isn't any way to actually access runpod deployed UI via a browser right?
Yes exactly. The purpose of this project is to have a headless ComfyUI API, that you can use via an API to run your workflows. What we have described via Local API is only something that should be used during development. But you could also use that with the deployed API on RunPod. If you just want to run ComfyUI on RunPod, it might be a better idea to use a Pod and use one of the templates.
Is the methodology described in "How to get the workflow from ComfyUI?" the only current supported method?
Yes exactly. There might also be an API in ComfyUI that provides a way to get the workflow, but I couldn't find any info about this. But maybe there is something, but we are just not aware of this.
When is says "Open ComfyUI in the browser", it's implied that you are running a separate ComfyUI server somewhere(locally perhaps).
When do you see this? When usingg the method described in Local API?
Can I run the UI locally, and have it render images on the runpod worker automatically when I press the queue button?
Or do I need to copy paste the workflow json each time I change something?
When you do Local API, you can run ComfyUI locally and also access the WebUI. When you use the WebUI, you don't have to send the workflow via the API yourself, the WebUI will already handle this for you. Both the WebUI and the runpod-worker-comfy are using the API provided by ComfyUI.
I'm a bit unclear about what the expected UX is for this application.
There isn't any way to actually access runpod deployed UI via a browser right?
Is the methodology described in "How to get the workflow from ComfyUI?" the only current supported method?
When is says "Open ComfyUI in the browser", it's implied that you are running a separate ComfyUI server somewhere(locally perhaps).
Can I run the UI locally, and have it render images on the runpod worker automatically when I press the queue button?
Or do I need to copy paste the workflow json each time I change something?
Related: #26
The text was updated successfully, but these errors were encountered: