Up next in 10
ComfyUI Desktop Unlimited FREE Forever AI – Run any AI Model to Generate Images, Video & Audio
Aug 22, 2025
Get the full source code of application here:
Show More Show Less View Video Transcript
0:00
Uh hello guys, welcome to this video. So
0:02
in this video I'll show you a opensource
0:06
desktop software or tool you may say.
0:10
With the help of this tool you will be
0:12
able to run any open-source AI model
0:15
directly inside your local machine. So
0:18
the tool name is comfy coom mfi.org.
0:22
If you visit this website, this is an
0:25
open-source tool, free open-source.
0:28
And as you can see, it's 100% free and
0:30
open source and no subscription, no
0:32
hidden cost. This tool allows us to run
0:36
any open-source AI model directly inside
0:39
your Windows machine or desktop, Mac
0:42
operating system. It's crossplatform
0:44
software.
0:46
So you can see all these images are
0:49
created using this software. So you have
0:52
this download button as you can see it's
0:55
also available on GitHub 75,000 star
0:58
server. You simply download this for
1:00
Windows, Mac or you can directly install
1:03
it from GitHub. For this to work you do
1:06
need to have a dedicated GPU card. Uh I
1:09
already have a Nvidia graphics card. So
1:12
this is must here. If you do does not
1:14
have a GPU then you can't run this. So
1:17
I've already installed it. It's a simple
1:19
step-by-step process and this is the
1:22
actual exe file. You simply run this and
1:24
this will open like this and by default
1:27
it comes with a single model installed
1:30
which is a stable diffusion which is a
1:32
very popular model for text to image. So
1:36
now as you open this application this is
1:38
a simple workflow which is actually
1:42
shown right here. So this is actually
1:43
constructed using nodes. This is the
1:46
first node where the model is loaded.
1:48
This is the second node where you
1:50
actually write the prompt here. Two guys
1:52
doing coding on laptop. And as we run
1:55
this model through our GPU, all these
1:58
parameters you can change. And this is
2:00
your output image which is shown right
2:02
here. So also this image if you want to
2:05
see this image you can open this image
2:07
directly in the browser. So this is you
2:09
can see the quality of the image and it
2:12
is running locally inside your system as
2:14
you can see local host at port number
2:17
8,000. So you can even directly open
2:20
this right localhost 8,000. So this will
2:24
open directly inside your browser. So as
2:27
you can see we are running this AI model
2:29
stable diffusion directly inside our
2:32
machine. So here we can change the
2:34
prompt here. So to other prompt here
2:37
let's suppose to inside two guys doing I
2:40
can simply say two girls
2:43
dancing
2:47
with
2:49
each other. So I basically changed the
2:52
prompt here and we basically get this
2:54
run button to run this AI model. So now
2:57
what happens if you see that now this
3:00
progress bar will show
3:03
instantly. Now using the tech text
3:05
prompt it will generate the image for us
3:08
and this inside this you can save this
3:11
image inside your local machine as well.
3:14
So automatically this image will be
3:16
saved as soon as it is generated. You
3:18
will see that. So this is a very handy
3:21
tool by which you can actually run AI
3:23
tools directly inside your local
3:25
machine. So now if you see the image
3:28
will change
3:30
and you can actually open this save this
3:33
image.
3:35
So
3:37
and similarly if you want to install any
3:40
other models you simply go to the model
3:42
library and here you can by default the
3:45
single model comes pre-installed but you
3:48
can even install more models by going to
3:51
the template section and right here we
3:54
have different templates out there which
3:56
is uh image generation image to image
4:00
as you can see we have uh all other
4:03
models which is flux
4:05
which is also open source right here. We
4:08
also have uh
4:10
you can even search for your own model
4:12
as well. This is your quen model which
4:14
is a Chinese-based model for image
4:16
editing. You simply click on this any
4:18
particular model and then it will prompt
4:20
you to download this model inside your
4:23
machine. So as this is not installed you
4:25
it is giving me these download buttons.
4:28
So you can download this model directly
4:31
inside your machine by clicking the
4:32
download button. And once it is
4:35
installed, you will be able to run this
4:36
model like this. As you can see, so it
4:40
has various templates automatically
4:42
created for you for each model. So you
4:44
simply click on a particular model and
4:47
automatically download this by clicking
4:50
the download button and automatically
4:52
the workflow is automatically created
4:54
for you. You simply change the prompt
4:56
and automatically your output
4:59
will be created. So these are
5:02
constructed using workflows
5:04
connected through nodes. So
5:06
automatically it is created for you just
5:08
need to download the model. After that
5:10
uh
5:14
you simply go to the template section.
5:16
You can also have different categories.
5:18
So this is for video generation. So here
5:21
you can install any video models for
5:24
video generation. You can use the latest
5:26
model model Chinese model 2.2. 2 and for
5:30
this you do need to install this. So
5:33
simply click the download button for
5:34
each model to install this.
5:39
So as soon as you install all these
5:41
models this will be automatically be
5:43
constructed for you. You will see that.
5:47
So here you simply write the prompt
5:50
and automatically your video will be
5:54
generated. So this is a very helpful
5:57
tool by which you can run any
6:01
open-source model directly inside your
6:03
machine. You do need to have a dedicated
6:05
GPU installed.
6:07
In my case, I'm using Nvidia GPU. So it
6:11
can be any GPU to run these models
6:14
directly inside your machine.
6:16
So it's very easy. You go to any sort of
6:19
category. If you want to generate audio,
6:21
you simply say audio
6:23
image API. So all these open-source
6:27
models are available. You simply click
6:29
in any of individual models by going to
6:32
templates
6:34
and you can run any model. For example,
6:40
this is the open AI model right here.
6:49
Click the run button like this.
6:53
So definitely check out this tool guys.
6:55
Simply download this and start running
6:57
any AI model directly inside your
6:59
machine. Comfy UI. It's also available
7:02
on GitHub. It's totally open source as
7:05
you can see here.
7:09
Comfy UI it is called right here. So you
7:12
can just check out their source code as
7:14
well. So it basically visually you can
7:17
run AI models through nodes. You don't
7:20
need to execute any command. So it's a
7:22
graphical user interface kind of a tool
7:25
where you exactly see what sort of
7:28
things are happening under the hood.
7:31
You constructed using nodes, different
7:33
nodes you connect with each other using
7:35
these connections that you see right
7:37
here.
7:41
So definitely check out this tool.
#Programming
#Open Source