video,

I'm going to be vibe coding with the newly released cursor

cloud agents.

And if you guys did not hear about this or have not tried

this out yourself yet,

I would highly recommend that you check this out.

And if you don't want to check it out,

if you don't want to pay for cursor,

that's one of the reasons why I'm making this video so that

you can actually see what's going on.

Because this in a lot of ways,

I think is the way that vibe coding is headed.

And I wanted to make this video just to cover this and to

talk about it.

And then to actually show you guys how it works,

because I think that it has a lot of potential.

So with that being said, there is a three minute video.

I'm not going to play this being go to just look up cursor

cloud agents and you can go to this website here.

I'll link it in the description down below as well.

But it says cloud agents leverage the same agent

fundamentals,

but run in isolated environments in the cloud instead of on

your local machine.

Okay, so this is very interesting.

Basically,

why would you use cloud agents over your local machine?

The reason is because they have access to their own virtual

machine.

Cloud agents can build,

test and interact with the change software.

They can also use computers to control the desktop and

computer.

So if we go over to this,

I just want to kind of show you guys what I was testing out

a little bit and show you an example right off of the start

here.

So check this out.

So I said at a social page for bridge mind where all

members of the bridge mind community can enter their name

and socials for recognition.

So basically what it does is it operates exactly like

cursor would on your local machine.

But instead of you then pushing changes to GitHub,

it integrates fully with Git.

And then it's going to, you know, hey,

you can see here bridge mind social page.

Here's the draft.

It's a draft PR and you can basically see the diff.

You can review it and then you can commit it as well and

you can click mark as ready.

So we're going to take a look at this because what it does

is it's actually able to run in a virtual environment and

then it literally will test the newly created feature that

you created.

Now it literally spelled bridge mind wrong,

but I want to show this to you guys because I think it's

important.

So it literally says viewing social page.

So this was the page that it created.

You can see it's at local host 3000 social up there.

And then it literally went through the process of entering

all of the different information into the form fields to be

able to test the functionality that it built.

Okay.

So once it tests the functionality that it built,

it basically says, okay, like it works.

Here's why.

And if we just skip through this, you know,

we can see that, okay, at the end it says,

thanks for joining, right?

So the feature that it built worked and it literally said,

Hey, this is me demonstrating the form submission.

Okay.

So this is definitely something that is different than what

we've seen from previous vibe coding tools and probably the

way that things are headed.

And I'll kind of get into that a little bit and I'll vibe

code with it with another task.

This was a task that I just did the other day on stream

with NS and he gave it to it.

And you know, we've kind of tested it out there,

but if we go here,

you can access this by kicking off cloud agents from

wherever you work.

So you can use this in cursor desktop, cursor, web, Slack,

GitHub, linear, or the API.

So literally on the desktop,

you can literally just select cloud and the dropdown under

the agent input and it'll be able to run this in the cloud.

Okay.

So if I pull up cursor,

what I want to show you guys is okay.

Let's like go over here.

Right.

And let's go to this and where even is cloud?

So it says select cloud from the dropdown.

I don't see.

Okay.

Here it is.

So work tree, I see work tree.

Do I need to refresh this or update?

Let's see.

Do I need to reload this and test this?

Where does it say to use it on the desktop?

Cause that would be useful to not have to actually access

it because in order to access it,

I actually have to go to cursor.com slash agents.

And then once you're here, this is how you're accessing it,

right?

So you're kind of reviewing it here.

You have full access here.

But my question is,

is how could we actually access directly from cursor,

right?

So it said what?

Okay.

There it is.

So I literally refreshed and now you can see that this,

so this is brand new.

I just had to reload cursor,

I updated it and then I had to reload it.

And now you can see that cloud is an option when you touch

the dropdown there.

So you have local work tree in cloud.

So let's click cloud here.

And I'm assuming that this is just automatically connected

to, I mean, the only issue that I have here is with cloud.

I don't necessarily have this connected to a branch, right?

Like I'm literally, if you guys look,

I'm in a mono repo right now.

So I wonder what'll happen here.

This will just be a quick test.

I don't know if this will work,

but let's drop in the bird's mind UI.

And right now I'm using,

let's switch over to a different model.

Oh wow.

So literally you can only use these models.

So you can only use GPT 5.3 codecs, high composer 1.5,

Opus 4.6, and then GPT 5.2.

So let's use composer one for this.

And I don't want to use, no,

I do not want to use multiple models.

Let's just use this one and I'm going to give it a

following prompt.

Um, okay.

Let's just say, hold on.

I need to get my, let's get bridge voice launch real quick.

It's actually, it's not launched yet.

So let's launch bridge voice.

Uh, if you guys don't know what bridge voice is,

bridge voice is the voice to text tool that I use for

bridge mind.

And, um, it's a phenomenal tool.

Um, I use it every single day.

It's faster than whisper flow.

It's a little bit more reliable.

It runs fully locally on your machine,

but I just logged in.

So now I do have bridge voice pulled up.

So if you guys haven't checked that out,

you can start your free trial today, um, at bridge mind.ai.

So now I have bridge voice and I'm just going to give it

the following prompt.

I want you to review this project and I want you to create

a new page called bug bounty event.

And I want it to have information about the bug bounty

event,

make sure that the page is styled according to the bridge

mind theme and other pages,

but this needs to be a page about information,

about the upcoming bridge mind bug bounty event that starts

on March 1st.

Um,

this bug bounty event will reward people for submitting bugs

about bridge mind and more information will be coming

shortly.

Winners will get big Bitcoin for bugs reported.

Okay.

So let's, um, let's just submit this and see what it does.

So it literally says push local branch to remote.

The branch main is not on the remote repository pushed to

remote before kicking off the agent.

So it says push branch to remote.

Um, the only issue that I don't know about this is like,

you guys can see that I'm in the mono repo here.

So I don't want to mess anything up because that's just

like that, that shouldn't be pushed to main.

So let's do one thing real quick.

Let's actually just see the into bridge mind UI and open

that up in cursor.

Cause that may be a little bit different.

So let's open this up in cursor and let's now pull this up.

And what are we on here?

We're on, we're on the dev branch right now.

So let's just run a get poll right now.

Um, so that we pull, let's do get fetch,

just make sure everything's up to date.

Um, okay.

So everything's good.

So now what I want to do is I'm just going to submit that

same prompt, right?

But now, um, we're going to switch over to cloud again,

and this should work.

I just want to make sure I don't mess anything up.

Um, let's zoom in a little bit here.

This is a good test in my opinion.

So we'll see what this does.

So now it's submits and now this time,

so you can see that that functioned a lot different because

I believe that this repo is already connected with my

cursor cloud.

So you can see there that it didn't even prompt me to set

up my environment, right?

Now it's just saying setting up environment.

So we'll see what this does and we'll kind of have this

running in the background,

but this is definitely something that, Hey,

you want to check this out because Hey,

this is in the cloud now, right?

Like this is cloud.

Uh, this is not local.

So we'll see what that does.

We'll come back to it.

But somebody said, uh, when I was live on stream,

that sometimes setting up the environment can take like 10

to 15 minutes.

So that's one thing to note.

I did also experience that when I was working with it as

well, but let's, uh,

let me pull this up and let's actually get back over to a

new agent real quick and oh wow.

So you can actually see, check this out chat.

So you can now see here that this prompt that I submitted

is now in the cloud in the dashboard.

So you can see bridge mind, bridge mind UI.

And that means that it did properly submit that branch to

the, you know, the dashboard here.

And it already is connected to that get hub repository,

right?

So we'll see what it does when setting up the environment.

I don't know how I necessarily feel about that,

but what some people were saying is that they were saying,

Hey, you know, I basically run this,

it's a long running task and I come back to it in the

morning and it's like, tested all these features, right?

So definitely an important thing to note.

Um, sorry, excuse me.

Um, so let's go over to new agent real quick.

And one thing I do want to touch on is like,

I had this go through and I had it go through my ARIA

labels to see like, okay, in terms of, Hey,

like can we improve SEO?

It came up with all of this and then it came up with this.

And what did it do in terms of like, get here?

I want to check that out.

So it said SEO draft ARIA labels,

and then here's the draft, right?

So can I just, is this already committed?

Like, did this push it to the main branch?

Um, let's ask real quick, where did you push this with get,

did you push it to the master branch or did you push it at

all?

Or did you create a PR or what?

Yeah,

I definitely want to see this because if I mark this as

ready, does it push it is let's hold, hold on.

I, we're just gonna, okay, squash and merge.

So, so what's squash and merge mean?

Squash and merge.

Let's just try this.

Um, let's just try squash and merge.

Maybe it means squash any merge conflicts.

Um, and it will do that automatically.

What is squash and merge mean?

Um, What does squash and merge mean in Cursor Cloud?

Let's see if we can figure that out,

because you guys see how it said squash and merge there?

Should you squash and merge?

Okay, so it says,

squash and merge is a method for combining multiple often

messy individual commits from a feature branch into a

single clean commit on the target main branch.

Okay, that makes sense.

Okay, so I wasn't familiar with that.

So now it does say merged.

So let's go over to the GitHub and see, okay,

did that actually deploy?

Let's see here.

Okay, let's go over to repos and okay, look at this.

So yeah, so it does say update now, updated now,

but what branch did it even update, right?

Did it update?

It did update master.

So you can see Matthew Miller and Cursor Agent SEO strategy

ARIA labels.

And here was that commit.

So we can go to this commit here and here's what it pushed,

right?

So this is what we're seeing.

It says coauthored by Cursor Agent,

coauthored by Matthew Miller,

and then it has all this information on it, right?

So this is actually really interesting to me because, hey,

this is showing you guys that, okay, like, hold on,

it did merge.

Let's not, hold on, let's stop here, stop, stop, stop.

Okay, so, you know, this is interesting.

You're able to run this fully in the cloud.

And one reason why I think that this is such a big deal is

because, hey,

we could get to a point where the AI agents are so good

that it doesn't even make sense to pull anything locally on

your computer for testing, right?

If you have the ability to be able to run it in its own

environment and be able to run these VMs,

then why would you need a developer to even be testing any

of this locally, right?

Okay, also, chat, it does look like this one is running.

So you can see editing bug bounty page client.txt.

So how long did it take for that to even get the

environment ready?

Let's see here.

So it looks like it took, I mean,

it took like maybe five minutes, I don't know,

to get that running,

but then you can see that it's creating the layout for it,

it's creating the page for it,

and then it's also creating the, it's updating the route.

So this is actually doing it perfectly.

So it literally went into my routes file like I needed,

now it's adding it to the footer.

Now there's another thing that we can do where, hey,

if we open this up here on the side,

you can see that this is the desktop that it's running.

So you can see that, hey,

these VMs with cursor cloud agents,

they have access to Google Chrome,

they have access to a terminal,

and then what even is that there?

So if we do take control, okay,

let's not interrupt the agent, but we'll see what it does.

We'll see if it basically like tests this,

but you can see like, hey, I mean,

this was able to test this,

is it was able to now run npm run build to make sure that

there's no build errors,

and now it's gonna be testing it in its own virtual

machine.

So this is interesting, this is different, and like I said,

this released the other day from cursor.

So if you're not up to date with this,

definitely important to take note of this.

It looks like npm run build completed successfully,

but while we wait for this guys to do the next moves,

I do want to say if you guys have not already liked and

subscribed, make sure you do so.

I do have a light goal of 200 likes on this video.

I'm pretty much live every day in our series of vibe coding

an app until I make a million dollars,

and the BridgeMind community is the fastest growing vibe

coding community on the internet with over 6,400

community members.

So if you guys haven't already joined the Discord

community,

make sure you do so there's gonna be a link in the

description as well as in the pinned comment.

So check this out guys,

you can now see it's doing git push origin cursor cloud.

Now we didn't see it test anything.

So if we go to git here,

it doesn't look like it's doing that quite yet.

It says that changes are committed and pushed to cursor new

cloud agent 3CE8.

Okay, so here is the git.

So now it's asking, hey, do you want to create a PR?

And now it created the PR and now it's saying, hey,

do you want to merge it basically?

Like once you say mark is ready.

Okay,

so one thing I'm interested in is why did it not test it on

the desktop?

Maybe you have to prompt it to do that.

Would you test this in the desktop to make sure that this

page looks correct?

Test it and make sure it's implemented correctly.

Okay, so let's test this and see, okay, did it,

did it actually, I just want to get a good,

one good example of setting this up on the desktop and I'll

also kick off another one.

So you can see, okay, it's doing playwright.

It's checking out the,

why is it checking out playwright config?

So editing bug bounty, I'm more so, what is it doing?

So it's saying playwright test.

So what is this doing?

So it just created an end-to-end test.

So you can see, look at this.

So it's using playwright and is that using playwright to do

all this?

Is it gonna,

I'm just curious if this is gonna show up here.

So now it's installing Chromium.

This is really interesting chat.

So what is it doing?

So it's, it's,

I thought it was just going to like launch it in the

browser here, like it did on that other example,

and then be able to show me like, okay,

this is me navigating the site and this is what it looks

like, but it looks like it's creating an end to end test.

Now, the question is,

is this what it looks like for it to actually then

implement that?

That's it.

That's what I want to figure out.

But while this loads,

what I want to do is actually want to kick off another

cursor agent in the cloud.

And what this one is going to do is it's going to actually

replace.

So if I do Burj my UI,

I want you to review the bug bounty page and I then want

you to update the Vibathon banner to become a bug bounty

banner and this new banner should have the information

about the bug bounty event and note that it starts on March

1st and also have a link so that users can view more

information about the bug bounty event.

So we're going to update.

There is currently a Vibathon banner, right?

So for our Vibathon, the Vibathon is now over.

So if you guys didn't know about this,

so if you go to bridgemind.ai,

you can see Vibathon 2026 and you click view results.

Here are the results for the 2026 Vibathon, Dr.

Vibe with first place consortium second and snitch with

third.

So this one just finished and what you can see now too,

and I just noticed this, so check this out.

So when you're in the cloud mode, not only can you,

you know, interface with this,

if you actually go to the website and you can see this

right where it's like, hey,

you can have access to get and you're able to mark as

ready, but you actually have a very,

very similar interface inside of the cursor web app itself.

So check this out.

So you can literally say, Hey,

view PR and Mark is ready for review.

So I can click view PR.

It's actually going to go to get hub and open that PR.

So it says new bug bounty banner.

You can review the PR and then you can go back over to

cursor and you can even say Mark ready for review.

And once you do Mark ready for review,

I think it's probably going to give the same thing.

So pull requests marked ready for review.

And then what can we even do here?

Can I, can I commit it from here or what do I need to do?

Oops.

Let's see here.

Go away.

Go away.

Go away.

Commit changes, create a checkpoint.

Okay.

Let's do cancel.

Okay.

So how can I, okay.

So here it is.

It gives you the same options that we had in the,

in the desktop version as we had in the web version.

So we can literally just do squash and merge.

And once we do that,

it's literally just going to push this to the master

branch.

So, you know,

I think that there's a lot that can be said about this,

but one thing that I do kind of just want to highlight is

that, Hey, this is, this is different.

Okay.

This is like definitely something that I'm personally going

to be using in my vibe coding workflow.

One thing that I don't necessarily know about is, Hey,

I mean, it did say, Hey, it created this end to end spec.

We didn't get to see it test it in the browser.

So I don't know why it didn't test in the browser.

Let's try that real quick.

I think that's a good way to be able to test that.

So if just ask it,

I want you to test this in the desktop and be able to show

me a recording of how this looks.

So can you make it so that, Hey,

you can just go and have it tested in the desktop similar

to like how anti-gravity does it, right?

It's like anti-gravity.

What's very different about anti-gravity in one way that

I've always said, okay, this is different, right?

Anti-gravity is different in this way.

And the way that anti-gravity is different than cursor is

that it does have that functionality where you're able to

see the anti-gravity browser open up and it has visuals of

the desktop and it's able to test it and click around.

But what we're now seeing is like, Hey,

if you go back to that example that I showed you at the

start of this video of what was it?

It was like the, it wasn't,

it wasn't the SEO or it was the website SEO.

It was the, hold on, let me find this real quick dashboard.

Hold on.

So it was the one that I showed you guys at the start of

the stream.

I can't find it now.

Okay.

Here it is.

Yes.

Okay.

Here it is.

Yeah.

So like,

how do you make it so that it does the walkthrough?

Like, do you have to prompt it to do a walkthrough?

Because it even shows you like, Hey,

here's what it looks like.

Here are the screenshots, right?

So I like this,

but why is there not an easy way for me to just say, Hey,

I want you to give me a walkthrough with this,

at least just so I can review it, right?

So I'm interested.

Like I thought this would be opening up and I guess this is

saying it's taking screenshots and it's probably going to

show me those in it,

but let's just wait here and see what it actually comes

back with.

So it says video on.

So it looks like it's getting ready for a video and we'll

probably just be able to come back to this.

So one thing that I'm noticing is that the functionality

related to the desktop and being able to get those videos

and to actually be able to review what it's created,

it's a little bit slow.

Now you can say that, Hey,

they're probably going to be able to speed this up for

sure, right?

But the thing is, if you're using anti-gravity,

anti-gravity pretty much does this immediately.

Whereas to spin up these VMs and look at this,

it's in its own isolated environment.

So this one's installing chromium again to be able to do

that.

So in order to even get like the ability so that you can

actually have those videos, it does take a little bit long,

right?

It's like, Hey,

if we go back over to this one where we are creating the

bug bounty test or the bug bounty page in the first place,

this one is still running and it's still like taking

screenshots and stuff, right?

So what I will say is that this is not,

this doesn't really feel like I'm going to be able to work

with this and have it be quick, right?

So if I'm working in the cloud and I'm working on this and

I'm saying, okay,

like I want you to show me what this looks like.

The issue with this is that it's not fast, right?

So if you're trying to make this so that people are able to

develop just completely in the cloud,

they're going to have to speed this up because yes,

I could see that eventually if this is fast enough,

if they get this to the point where this is almost

immediate and this doesn't take like 10 minutes to be able

to produce a couple of screenshots or like, you know,

take the video and show me what the page looks like,

then absolutely.

I could see how this is like totally the future right now

though.

Hey, if I'm a developer and I'm developing locally,

I want to see the bug bounty page right now.

And the only way for me to do that is to,

to run npm run dev,

open up my local host and look at it myself.

Whereas with cloud agents, you know, what,

how long have we been waiting for this to even produce

anything, right?

It's been over, over 10 minutes, right?

I mean, it's, you know, been over at least 10 minutes.

So this is one thing I definitely am not saying we were

having troubles reaching the AI provider retrying

automatically.

So I don't know what we'd be able to even see this.

Let's go back over to the bug bounty page and see,

was it able to get this?

So it does say, okay, look at this.

So literally I asked it, I said, I want to see it, right?

And video in the workspace route.

Oh my gosh.

Well, can you show me it?

Can you show me the video in chat here?

So I wanted to, I wanted it to test the desktop.

Do I have to prompt it to test it on the desktop?

Would you test this on the desktop?

So this is very interesting.

So we're kind of like playing around here with it, but,

you know, it created a video in the browser,

dragging into Chrome, and it now wants me to, you know,

test it, and then it says run it locally.

So like, here's the thing is like, okay, like, yes,

it did do something, but now it's just telling me, hey,

like open this up, you need to run this yourself, right,

and I'm like, hey,

what I wanted it to do is I wanted it to test this on the

desktop.

So let's see if I prompt it,

would you test this on the desktop?

Will it be able to take control here and actually go to

that page?

That'll be a good test because, hey,

then it just means that you just have to test it yourself,

right?

Like just prompt it to test on the desktop and it'll be

able to do that.

But that's not currently what I'm seeing.

What we had to do previously is we had to make it so that,

you know,

you had to go and it's creating all these screenshots.

You know,

you can even see here that this one is still kind of airing

out.

So we have to give cursor a little bit of grace, right?

They just released this feature the other day,

but the reason I wanted to make this video is because cloud

agents and vibe coding in the cloud is going to be

something that's going to become more prominent in the

future.

You're not going to have developers developing locally on

their machine anymore in 10 years.

Most likely it's going to be an approach like this where,

hey, you have the desktop on the right.

You have agents that are fully set up with their

environments.

You're able to be able to configure their environments

because that's another thing that I wanted to show you.

And I don't want this video to get too long,

but another thing that I could show you guys is like, hey,

what does it look like when you're actually setting these

up, right?

So check this out.

So when I initially set these up, right?

And you can see my cloud agents running here.

I have two agents here, but check this out.

You can actually integrate this with any MCP that you want.

So if we wanted to connect the BridgeMind MCP, we could,

but just out of the box,

these are the MCPs that it lets you connect to.

But one thing I want to show you guys is let's say I wanted

to set this up with the BridgeMind API, right?

When you guys saw that setting up the environment,

what's really interesting about that is that you can

actually set up this environment, right?

So you go to this page here,

and this is how you actually set up your environment.

And this is what's really, I think,

different about cloud agents versus developing locally,

is that let's say that you have a huge team, right?

And your team uses cursor.

Well, if you're using cloud agents,

you just set up the workspace once.

You literally just paste all of your EMVs into the

environment here.

And this is very similar, this is not new.

We've seen Codex Cloud have this as well, as well as,

I think, Cloud Cloud, right?

You create these isolated environments and you enter your

EMVs, and then it's able to kick all of this off.

So this is something that I just want to highlight to you

guys that this is happening.

Well,

I'm probably gonna be using this a little bit more because,

hey, like I've always said,

this is probably something that is the future, right?

It's like, okay, check this out.

So I said, would you run this on the desktop?

And let's see if it was able to run this, and it wasn't.

So that's interesting.

It said, test this on the desktop.

It just wants me to go local.

Now let's go over here and see, did this deploy?

Okay, so check this out, chat.

So now we can, this is shipped.

So we did just push,

you can see this bug bounty event starts March 1st,

and you can click learn more.

And then this is the bug bounty event page that it created.

So what we're talking about here is we're talking about

very, very similar, like very simple tasks, right,

that we just gave it.

So I don't know how I feel about it, because, hey,

if I was developing this locally,

I probably could have gone through a couple iterations and

been able to like see this a little bit better on my end

personally as a developer, versus, hey,

I have to like push this to Git, and I have to deploy it,

and then to even see it.

You know, it's like, hey, they do have this desktop, but,

you know, like we saw,

it just takes a little bit of time to actually generate

those tests.

I mean, this one still isn't even,

this one just bugged out completely.

And then this one never showed us a video, right?

Even though I said, would you test this on the desktop?

It never did that for me.

It created a bug bounty banner webm file.

So this is something they'll probably work these kinks out,

but what it's supposed to do is it's supposed to do exactly

what it did inside of that social page that I showed you

guys where it literally will demonstrate the feature,

right?

So I'll be working with this a little bit more on stream

today.

I am definitely interested in how this is going to evolve,

because I think that, yes,

this is probably going to be something in the future that

you are going to want to use and be up to date on,

because developing in the cloud,

using cloud agents that are set up with their own custom

environments that run on isolated VMs,

this is gonna be something that is used a lot more in the

future.

You're not just gonna have developers developing locally

anymore.

There's gonna be vibe coding in the cloud.

And I think that this is a big step towards it.

So with that being said, I'm gonna end the video here.

If you guys haven't already liked,

subscribed and joined the discord, make sure you do so.

And with that being said,

I will see you guys in the future.