Up next in 10
VSCode + Cline + Continue | NEVER PAY for CURSOR again. Use this OPEN SOURCE & LOCAL Alternative
Apr 23, 2026
https://gist.github.com/gauti123456/aec9cf454c4aa6287c0bf20e0bf623ce
Show More Show Less View Video Transcript
0:01
Uh hello guys in this uh live stream I
0:03
will actually show you the complete
0:05
workflow inside your VS code by which
0:08
you can completely replace cursor.
0:10
Cursor as you all know it's a number one
0:12
AI tool for developing coding
0:14
applications. So I will show you the
0:17
complete free workflow by which you can
0:19
completely replace uh cursor inside your
0:22
VS code. And for this we need two
0:24
extensions. First of all, we will I will
0:27
show you how to set up a local AI
0:29
provider using an extension called as
0:32
continue. So, if you simply go to
0:34
extensions tab, simply search for this
0:36
extension. If you simply search
0:38
continue, it's an open-source AI coding
0:41
agent. And for this, you need to install
0:44
this. So, click the install button.
0:46
Simply install this extension. So, you
0:48
can see it's almost having 3 million
0:50
installs. I have already installed it.
0:53
And after that you also need ClintIN
0:56
extension as well. Clint is also a very
0:58
good extension coding extension for
1:00
actually integrating different APIs
1:03
directly inside your coding VS code.
1:06
Just install this extension as well. So
1:08
this extensions is also needed. So just
1:11
install these two extensions. After that
1:13
you also need a Olama. Olama is a
1:16
actually a software by which you can
1:18
download AI models directly inside your
1:21
machine. So simply download Oola inside
1:23
your machine. I have already downloaded
1:25
this. So as you see if you click the
1:27
model section, it actually contains all
1:30
the popular AI models.
1:33
It is there. All the open-source models
1:36
which are directly there. So once you
1:39
install this continue extension, this
1:41
continue extension will directly appear
1:43
right here. If you see, click the
1:46
continue extension right here. And in
1:49
the description of this live stream, I
1:51
have given this uh config. Um file. So
1:55
once you go to this extension, simply go
1:57
to the settings tab. Simply click on
2:00
that and right here you will find out
2:02
this config. Go to the config right
2:05
here. Just go to this and just type this
2:09
conf local config. Click this option
2:12
local configuration. So this is actually
2:14
a configuration file. So if you check
2:17
the description, I have actually pasted
2:19
this code right here. So simply copy
2:23
paste this overall code. This actually
2:25
is actually telling the continue
2:28
extension that we will be serving this
2:31
AI model which will be running directly
2:33
inside our local machine through Olama.
2:36
So I basically put these two models
2:39
which is first is Quen 2.5 coder
2:43
7 billion parameter and this one is quen
2:45
2.5 coder 1.5 billion parameter. So both
2:50
these models first of all I need to
2:51
download this through
2:54
and you can actually find this source
2:55
code I have given this in the
2:58
description of the live stream you can
3:00
actually check the full source code and
3:02
directly paste inside this file. After
3:05
that now I need to actually install both
3:08
these models. So I will simply
3:12
once you install
3:15
you will actually start the right here.
3:19
I have already downloaded both these
3:21
models. If you want to download these
3:23
models through command line it's very
3:24
easy. Go to command line. Simply type
3:29
and pull and then put the model name.
3:34
This is a command to actually download
3:35
this model inside your local machine.
3:39
Similarly do for the second model as
3:41
well. So you can directly search
3:44
if you simply write here coin 2.5.
3:48
So this is actually both these models
3:51
that I installed. You can take any
3:53
example depending upon your machine. So
3:56
you should have a powerful machine to
3:58
actually run higher AI models. So
4:01
depending upon your machine. So these
4:03
two models first of all I'm jo show
4:05
showing you this it actually the size is
4:08
4.7 GB and then this one is 986 mgaby so
4:13
this will work on all the machines so
4:16
just install these two models simply
4:20
run these two commands I have already
4:22
installed it
4:24
soap pull just write these two commands
4:27
to install this so once you install this
4:30
once you in uh execute this command. You
4:34
will see both these models will be
4:35
there. I've already installed it. So
4:38
this is the first model 4.7986.
4:42
So after that simply open
4:45
and this will be running. And now just
4:48
open your continue extension once again.
4:51
And after you pasted that code simply
4:54
chat with this send out the message hi
4:57
and now the actual model will be running
5:00
through Olama. Once you ask any kind of
5:02
question the response will be generated
5:10
just make sure that you have a good
5:11
powerful machine to actually run the
5:13
models. So it will take some time. So
5:16
now you can see it actually send out my
5:18
message hello how can I assist you? So I
5:21
will simply say please
5:24
make a index html file which uh is image
5:30
to PDF
5:32
editor in browser.
5:35
So I ask a question that please make a
5:37
index html file which is image to PDF
5:39
editor in browser. So now the AI will
5:42
start generating the code for me
5:44
directly.
5:45
So by this approach you can directly
5:49
locally install the models and directly
5:51
use it and completely you can replace
5:54
this cursor.com
5:56
website which cost you a lot of money.
5:59
So that's is the first approach. The
6:02
second approach by which you can
6:04
actually
6:06
integrate.
6:10
So let me just stop this.
6:14
The second approach by which you can do
6:16
that is through the extension which is
6:21
clean clean extension.
6:29
So
6:36
let me just open task manager and
6:38
directly. So uh you can see in my Nvidia
6:42
GPU is stuck at 100%. So when you
6:45
actually execute these locally models if
6:48
your machine is not powerful then it
6:50
will be hanged. So in that scenario you
6:52
need to actually close this app. This
6:56
can actually make problems. So the
6:59
second solution I will show you is
7:01
through actually running the models
7:03
through the API. So this API is called
7:07
as build.envidia.com. nvidia.com/models.
7:10
This website allows you to execute all
7:13
the models through the API. So simply
7:16
register on this website and once you
7:18
register this it will provide you with
7:20
free API key. So just verify your
7:23
number. I have already created the
7:24
account. So once you go to this simply
7:26
select any model that you want to. So
7:30
let's suppose I want to
7:34
select any kind of model. From here you
7:36
can select
7:39
let's suppose I say I want to use this
7:42
one Gemma 4. So I simply click on the
7:46
model that I want to use. After that uh
7:53
you click on this button view code.
7:55
Simply click on that and once you click
7:57
on the button you will be seeing the
8:00
instruction to actually integrate this
8:02
model.
8:10
So right here you will basically paste
8:14
this endpoint.
8:19
This is actually the endpoint that you
8:21
need to paste.
8:23
So just open your extension. Uh
8:35
so just open the clean extension.
8:41
So once you open this you go need to go
8:44
to the settings button to actually
8:46
configure it. So for configuring it go
8:49
to the settings button right here.
8:56
And from here in the API provider, you
8:59
simply select OpenAI compatible
9:03
and just paste this base URL that we I
9:07
showed you.
9:12
So after you pasted it, now you need to
9:14
get the API key. And for getting the API
9:17
key, it's very easy.
9:19
Uh you go to the profile section. Uh
9:22
from here you simply select your API
9:25
key.
9:31
So go to the account section and from
9:33
here
9:39
go to API key.
9:49
Click on generate API key and uh simply
9:52
give it a name
9:56
and after that click on generate key. So
9:59
after that it will give you the API key
10:01
that you want and simply copy this API
10:04
key
10:06
and uh
10:10
simply here paste your API key and from
10:15
here you replace the model name. So in
10:17
my case I'm using Gemma 4.
10:21
So after you do this click on the done
10:23
button. So now you have successfully
10:25
configured this and after that you
10:28
simply send out the message
10:32
please make a
10:36
image to PDF HTML file
10:40
which contains It's
10:57
so as soon as you send out this request,
10:59
the AI will start developing your
11:01
application using the API that you
11:03
configured. And uh this is the second
11:08
approach by which you can directly
11:10
integrate any AI model. So this gives
11:13
you free API request. So if you have a
11:17
powerful machine, you can follow the
11:19
first approach. If you don't have a
11:20
powerful machine, you can create an API
11:22
key and directly configure through Clint
11:26
plug-in which is really.
11:32
So now you can see the a AI will start
11:37
developing your application.
11:39
In this way you can completely replace
11:41
cursor
11:44
through these two approaches that I
11:46
showed you.
12:28
So it will start creating the
12:30
applications for you. So after you send
12:34
out this request