Jun Li

And they have both open source and closed source.

So closed source is essentially when you need an API.

These are the larger models where you have to make a call to their service to be able to run it.

My focus as, of course, because of the LM Studio example is I want to run it locally.

So there's a couple of different, I should say, you have many different options, right?

Most of this is text to text.

video, image to video, image to image, and you can run any of these on your system.

Because we've taken a look at Higgs field in the past, Higgs field and those kinds of services essentially are a packaged cloud-based service for doing this kind of thing.

So if you've ever wanted to run it locally or be in control of the entire generation process, comfy UI is going to be that kind of, um, source for you as opposed to, um, you know, like a Higgs field or a galaxy or, you know, any number of the ones that exist.

Um, right.

Absolutely, absolutely.

And so when you go through this process and you say, hey, I want to try a new text-to-image video and whatnot,

you can just simply select the model and there's this nice but somewhat confusing process here.

But

Essentially, they have their own interface process.

And it says, when you click on the template, it's trying to look for the models that are associated with the template.

So you have to go through processing by downloading each of these models.

Once they're loaded, and you can do it more directly, and this is an automatic feature, which I found helpful when first starting using it and things like that.

Remember, this is all open source.

This is all connected to Hugging Face and GitHub and that whole infrastructure.