Up next in 10
🧠 1. Multi-Model Platforms (One API → Many Models)
OpenRouter
https://openrouter.ai/
Together AI
https://www.together.ai/
Hypereal AI
https://hypereal.cloud/
⚡ 2. BEST “REAL FREE” APIs (No Card Required)
Google AI Studio
https://aistudio.google.com/
Groq
https://console.groq.com/
Cloudflare Workers AI
Show More Show Less View Video Transcript
0:01
Uh hello guys in this uh live stream I
0:04
will actually show you how to actually
0:05
use the latest Google AI model which is
0:09
called as Gemma 4. Gemma 4 is actually
0:12
released by Google recently. This is
0:15
their latest AI model. So I will show
0:17
you how to use this model directly
0:19
inside VS code for completely free. So
0:22
there is a website which offers you free
0:24
API key that you can actually use it. So
0:27
f first of all you what but what what
0:29
you need to do you need to install an
0:30
extension in VS code which allows you to
0:33
directly use this GMA 4 or any kind of
0:36
AI model directly inside your VS code.
0:38
So right here you will see clin either
0:42
you can install ru code or clin both
0:45
these extensions will work but I will
0:48
prefer this extension clin simply click
0:50
the install button just activate this
0:52
extension I have already installed it.
0:54
So once it appears it will appear right
0:57
here like lane. So just activate this
1:00
extension. So after you activate this
1:02
you will actually
1:05
uh that extension will start. So now you
1:09
you need to go to settings option. So
1:11
right here you will see the settings
1:12
option. Simply click on that and right
1:15
here you simply select from the open AI
1:18
provider. Inside the API provider right
1:20
here you need to select open AI
1:23
compatible. So just select this option
1:25
from here open AI compatible.
1:28
And the website that we will be using
1:30
for getting the API key it's uh Nvidia
1:36
and build.envidia.com.
1:39
So simply go to the website. So this
1:41
website provides you with free API keys
1:44
once you sign up completely on this
1:46
website. I already done that. So it
1:48
allows you to simply sign up using your
1:51
Google account. So once you sign up, you
1:53
actually get these API limit 40 request
1:57
per minute you can send out to these AI
2:00
models. So once you go to the model
2:03
section, you will be able to find all
2:05
the models right here. So all the
2:07
popular models are there. You can almost
2:10
191 models are there. So from here you
2:13
can search for your favorite model. If I
2:15
simply search here,
2:18
I simply search Gemma right here. So you
2:21
will see this model will come right here
2:23
which is by Google.
2:25
So all the famous models are there. So
2:28
right whenever you sign up you simply go
2:30
to the API key section from here. Simply
2:33
click this and right here I have already
2:36
created a API key. So what you can do
2:39
you can delete. Let me delete this and
2:41
create this once again.
2:44
So right here you see this option. Click
2:46
on generate API key. Simply name
2:48
anything. Just give it a name and from
2:50
here expiration you can even control. So
2:53
by default it's set to 12 month. Click
2:55
on generate key. So after you generate
2:58
the key your API key will be created.
3:00
Simply copy to clip import and just
3:02
close it. And from here you simply paste
3:07
this API key from here. Simply paste it.
3:12
So this provider the base URL
3:17
you need to write like this simply https
3:21
integrate.api
3:24
nvidia.com/ver1.
3:27
So you can even do this as well if you
3:29
go to the explore section or model
3:32
section.
3:34
So simply select any model. For example
3:36
you are selecting this model.
3:41
Just select any any of the model that
3:43
you will be integrating. So let me
3:45
search for Gemma 4.
3:48
So let's suppose you'll be want to
3:50
integrate this model. I simply click
3:52
this and it will give you the
3:55
documentation for this. So once you
3:57
click this uh you simply
4:01
in the right hand corner of the screen
4:02
you see this option view code. Simply
4:04
click this and right here you it
4:07
actually gives you this URL. This is the
4:09
actual the base URL that you need to
4:11
paste just this portion. So simply copy
4:15
this portion from here. Simply Ctrl C
4:18
and now go to that plug-in. Simply paste
4:21
inside the base URL and just pasted your
4:24
API key. This is the base URL and from
4:27
the API provider you select OpenAI
4:29
compatible. After that inside the model
4:32
whatever model that you will be
4:34
targeting. So in my case we will be
4:36
targeting this model which is so simply
4:40
copy this exact name simply right click
4:42
copy and directly paste it. So very
4:45
straightforward process you simply set
4:48
up this after that you simply click the
4:52
done button. After you configured this
4:54
now simply restart
4:57
uh simply click start new task and we
5:01
selected this model as you see open AAI
5:04
compatible Google I simply ask a
5:06
question I send out the request hi so
5:09
now the actual model will respond to us
5:12
this clin it will be using that API key
5:15
that we configured from Nvidia and for
5:18
the very first time it takes around
5:20
about 5 to 10 seconds to actually
5:22
respond to the request and after that
5:25
subsequent request will be faster.
5:29
So this actually is a free website
5:31
Nvidia
5:33
is saying that uh status code nobody 404
5:37
model ID.
5:41
Uh I think
5:44
uh the model ID is incorrect. Let me
5:48
just try this. paste it.
5:57
Let me try this with a different model.
5:59
For example,
6:02
it is very simple. You can actually let
6:05
me select this one which is GLM 5.1
6:14
or I think this one
6:18
free endp point
6:21
some models will be unavailable as well
6:24
some models so which are having free
6:27
endpoint So
6:39
as you can see some models are
6:41
downloadable some are free endpoint. So
6:45
whichever is having this free endpoint
6:47
they can also be used as a
6:51
so
6:54
from here you simply select your model.
6:56
So I think I can use that Kimmy model
7:00
which is really pop popular model. So
7:12
so we can select this model for example
7:15
and click on view code. The
7:18
the previous request was not successful
7:21
because many people are using this
7:23
website. So that's why some of the time
7:26
the model will not be responding. Simply
7:29
paste this one
7:31
and click on done and again retry.
7:36
So as this is a free website sometimes
7:40
you will get this error. provider ID. I
7:44
think I provider ID needs to be
7:51
open AI.
8:00
So right here you select
8:04
I think that needs to be
8:12
Open
8:25
the process is same guys. Sometimes due
8:28
to the
8:32
due to the website sometimes it can
8:35
create problems
8:37
but majority of the time the this will
8:39
work. it will return out to you the
8:41
response.
8:47
So the process remains the same. You
8:49
create the API key, simply paste it
8:51
inside Clint. If this plug-in is not
8:54
working, you can try R code which can
8:58
also work.
9:07
So let me try this with a different one.
9:10
Root code. I also have root code also
9:13
installed. Same course settings will be
9:15
same. From here you select your open AI
9:19
compatible.
9:21
the base URL. You simply
9:34
think I putting this base URL.
9:38
I think that was the issue. We putting
9:40
that base URL incorrectly. Let me go to
9:43
that clin.
9:47
The base URL needs to include also
9:49
version one. That was the problem. Let
9:51
me paste the correct one. This needs to
9:54
be version one is also included. That
9:56
why I I think that was the problem. If I
9:59
retry this, you will now see the request
10:02
will be successful.
10:05
So now you can see that guys the request
10:07
is successful. Now it was just the
10:10
problem
10:12
with the base URL. Now you can see it is
10:15
asking you questions. Hello, I have
10:18
greeted me.
10:21
So
10:23
I can simply say build a image to PDF
10:28
web app in browser. Create
10:31
the index html file. So simply I ask a
10:35
question build image to PDF web app in
10:37
browser. Create the index html file. So
10:40
now it will create a fully full stack
10:43
coding application for us. So it will
10:45
create the file for me. You can see that
10:48
in the right hand side corner of the
10:50
screen. It is writing the code for us.
10:53
So this is a full setup process guys. I
10:56
showed you step by step how to set up
10:58
this inside your clin. We are using the
11:01
free API key provided by Nvidia which is
11:05
actually
11:06
running all the AI models. You can run
11:09
Gemma, Kim, K2, any kind of model.
11:14
So you can see it is writing the code
11:15
for us.
11:18
So once it completes you will get a
11:19
notification that it is completed.
11:24
The speed is also very good as you see
11:26
that the speed of generating the code.
11:29
So this is actually the extension which
11:31
is really popular in VS code
11:35
which actually write the code for you.
11:44
So once it completes it will actually
11:46
show you the file is completed either
11:49
you can save or reject the changes.
11:52
I click on save.
11:54
So now I can run this application as
11:57
well.
12:00
So
12:02
now you can see it has built out a
12:04
complete image to PDF converter
12:06
application. It's completely responsive
12:09
as you can see that. So it actually
12:12
generated all this description as well.
12:15
So you can see that
12:19
it also runs run your application as
12:21
well. So it basically also run the
12:25
application as well. So this is the
12:28
process by which you can run any kind of
12:30
model. It can be JMA 4 as well. So
12:33
instead of Kimik K2 in moonshot, you
12:35
simply paste whatever model that you
12:37
want to run. So in our case, if you want
12:39
to run Gemma 4, you simply select uh the
12:44
model from here. Simply
12:49
in the search bar, simply select the
12:51
model that you want to use.
12:58
So all the free models are there. You
13:00
also have quen model which is also a
13:03
very good model, Chinese-based model. So
13:06
in my case I want to use Gemma. So I
13:09
simply search this.
13:14
So we just replace the name of the model
13:17
from here. So very simple. So in this
13:20
easy way
13:22
the process remains the same. You paste
13:25
the name of the model from here. Simply
13:27
click on done.
13:31
Uh I can simply say please build a
13:36
PDF editor web app in browser.
13:41
Create a PDF editor HTML file. So now
13:46
you will see that it will now again
13:48
start a new application PDF editor. It
13:51
will start writing the code for us on
13:53
the right hand side. This time we are
13:56
using this uh Gemma 4 model not Kim K2.
14:04
So
14:06
uh this is my personal favorite plug-in
14:09
uh Clint and I use this Nvidia which
14:12
allows you to execute 40 request per
14:15
minute. Just create account just verify
14:18
your phone number and you will get uh
14:22
free API key for you. So you simply
14:24
create that and then
14:32
so we just simply ask the question how
14:34
the model is thinking about that.
14:43
So in this easy way guys you can
14:45
actually use this GMA 4 model directly
14:48
in VS code.
14:52
So,
14:57
thank you very much for watching this
14:58
video. Please hit that like button,
15:01
subscribe the channel as well.