Why 1M People Tried This AI Coding Tool (Full Vibe Coding Tutorial)

Why 1M People Tried This AI Coding Tool (Full Vibe Coding Tutorial)


Check out Emergent: https://app.emergent.sh/?via=1666c8

In this video, I'll be giving you an in-depth vibe coding tutorial where by the end of it, you'll know how to build large production scale applications simply by using a prompt or using natural human language. While you can do this without any development experience and this will be a beginner friendly tutorial, if you are a developer, this will be significantly easier and you will just be able to create something more impressive because you actually understand development and just bluntly, you know how to prompt the model better.

Want to make real money with coding? I share high-signal insights on careers, monetization, and leverage in my free newsletter. Join here and get my guide How to Make Money With Coding instantly: https://techwithtim.net/newsletter

🎞 Video Resources 🎞
Emergent Y-Combinator Page: https://www.ycombinator.com/companies/emergent
Emergent Prompt Guide: https://help.emergent.sh/articles/169224-guide-to-prompt
Emergent Features & Tools: https://help.emergent.sh/articles/169224-guide-to-prompt

⏳ Timestamps ⏳
00:00 | Overview
00:56 | What is Vibe Coding
03:00 | Vibe Coding Platform (Emergent)
04:22 | Crafting the Initial Prompt
10:30 | Web App Architecture Breakdown
12:32 | Live Preview
15:52 | Saving to GitHub (Checkpoints)
18:24 | Build Phase & Debugging
26:10 | Debugging Framework
27:46 | AI Integration
28:49 | Advanced Features (Fork Mode)
31:00 | Deployment Phase
32:40 | Credit Usage
33:20 | Additional Features

Hashtags
#emergentai #vibecoding #softwareengineer
Closed Caption:

In this video, I'll be giving you an in-depth vibe
coding tutorial
where by the end of it, you'll know how to build
large production scale applications
simply by using a prompt
or using natural human language.
Now, while you can do this without any
development experience and this will be a beginner
friendly tutorial, if you are a developer,
this will be significantly easier
and you will just be able to create something
more impressive because you actually understand
development and just bluntly,
you know how to prompt the model better.
If you're a complete beginner, don't worry.
What I'm going to aim to teach you
here is the best practices how to use these tools
most effectively,
and how you can actually create a full application.
So something that isn't just a front end,
for example, or a nice landing page,
but something that has authentication databases,
some security built into it, maybe an AI integration.
I'm not sure exactly
what the app will look like at this point,
but I want to make it advanced and kind of full scale
or full stack, right?
So we have front end backend databases, etc..
So with that said, let's quickly
define what vibe coding is.
Then I'll hop onto the computer
and I'll show you the platform we'll be using,
and we'll talk about kind of the architecture
of a vibe coded app.
So you understand the different components
that we're going to need to build.
All right. So first things first vibe coding.
What is it.
Well simply put this really just means
that you're using natural human language
to create an application.
So rather than dealing with frameworks and languages
and writing a ton of code and needing
to be a professional developer,
you simply tell an AI model what it is that you want.
Now, importantly, you need to know what you want,
which is something that we're going to talk
about more in this video,
but it's more like being a product manager,
someone kind of leading a team or
guiding a bunch of developers and indicating to them
what it is that you want built, what you want
changed, what features you want added, etc..
So there's a lot of ways
that this can go wrong, obviously.
And if you know the best practices,
you can get a significantly better result,
which is what I'm going to share with you here.
Now, this by no means replaces developers.
In fact, most of these vibe coded apps
need developers to maintain them.
All of the applications you use that enable vibe
coding are built by real developers.
You're not vibe coding a vibe coding application,
if that makes sense,
and there still is a lot of place for that.
So I know a lot of people get worried
when they see these videos, I wouldn't be
this is just the direction that development is going,
and real developers are still very much in demand,
as I've seen a lot recently.
Okay, with that said, let's talk about the platform
we're going to use now for vibe coding.
There's a ton of different options out there.
There's many different platforms.
Now for this video,
I'm going to be using a platform called a merchant.
It is free to use.
You can get some free credits.
It's a relatively new platform.
It's only been out a few months now,
and in just two months they had over a million users
sign up for it, and they had over
$10 million in annual recurring revenue,
which is a massive amount for such a new company.
It's also backed by Y Combinator,
and they kindly sponsored this video.
Don't worry, it is free to use, and a lot of stuff
I'm going to share with you here.
You can apply to other platforms as well.
It's just the one that I'm going to use here,
which I think is new cool
and that you guys should know about.
Anyways, that's the platform.
Now let's hop over to the platform
and start Vibe coding.
All right.
So I'm on the emergent website
here. We're going to go ahead and get started.
But I do want to talk a little bit about the pricing
because I know that's something
that's going to come up now.
This platform,
like many of them, are free to use and try out.
However, you will have limited credits and you may
not have access to all of the advanced features.
So in this case, what I've done is
I've just purchased the smallest subscription
they have, which is $20 per month,
which gives you access to some of
the more advanced features and some more credits.
You do not need to do that.
If you want to vibe code again, you can play around
with all kinds of platforms what you're going to see,
because these are quite expensive
to run in terms of the Lem calls,
is that mostly you can use them for
maybe an hour, hour and a half.
You can get something basic built, and then pretty
much all of them are going to charge you some money.
So I usually recommend, if you really want to learn
this, be willing to spend probably 20 bucks, right?
20 bucks a month.
That's usually what I'm paying
for most of these subscriptions.
And that's going to get you quite far.
A lot of them have, you know, $200 per month
or $1,000 per month subscriptions.
You don't need anything like that.
20 bucks is usually totally fine.
Okay.
So anyways, that's what I have right here, just
in case you see something, you know, on my screen
that's not on yours
is because I do have that pro subscription.
But again,
a lot of what I'll share will apply no matter what.
So the first thing you're going to do on immersion
or in any of these platforms
probably is just type in an initial prompt.
Now the initial prompt is very important,
and this is how we're going to kind of set
the context of our project
and tell the model what it is that we want to build.
So I've just opened a web page here, which I'll link
in the description, which is a kind of detailed
guide on how to actually prompt the model
and how to create an initial prompt.
Now, general best practices are
that whenever you're prompting these models,
you need to be very clear on what it is
that you actually want, right?
You need to be specific.
You need to give all of the details
and you need to say more.
So what you want rather than what you don't want.
A lot of people say, don't do this, don't
do this, don't do this.
That's not really the best way to go about it.
It's more so telling it directly what you do
want it to do.
So see here. Define clear business or personal goals.
So also rather than just telling it what you want,
tell it why you want that
and what the objective of this application is
so that it can build that context in.
So it shows you an example of an initial prompt here.
Right. Here's the business. Cool.
Here's some features that we need.
Here's the target users.
Here's some success criteria okay.
What I usually do whenever I'm prompting these models
is I go into like a voice mode or a dictation.
I dump it to like ChatGPT.
I just talk for like 2 or 3 minutes on my idea,
as if I was sharing it with a friend, for example.
And I tell it craft an effective prompt.
Based on these guidelines, I would just copy
the guidelines directly from this site
to something like ChatGPT
that I can pass to an AI model
to create an application, so I'm never writing
the prompt completely from scratch.
I'm usually just kind of dictating my thoughts
into an AI model
and telling it give me
an effective prompt based on that, and it does.
The organization and kind of,
you know, applies the best practices.
For example, be explicit and specific,
provide the context,
use pause
the framing and quality modifiers, for example.
So rather than create a dashboard, create
a comprehensive dashboard.
Right.
And then after that, in this specific tool
the agent will start creating a more in-depth plan.
That's something that happens in a lot of these
what is it vibe coding tools.
So I actually already created
a really in-depth prompt here.
I just have it in Google Docs.
Again, I thought through the application
that I wanted to build quite in-depth,
and I'm just going to copy this whole prompt.
I'll go over some components of it
with you in a second and paste it into a margin.
So let's just go ahead and paste this for right now
and get it started.
And then while it's working, I can explain to some of
the things that I did and what I'm trying to build.
It's very important that you know what you want
to build before you start here, because if you don't,
you're going to get all kinds of wild results
and the model is not really going to know what to do.
It's going to bring you in a lot of weird directions.
So I'm going to press run here, but quickly
I'll let you know that you can change
the agent type
that you want to use here to E1 or E1 point one.
If you want it to be fast, then change it to 1.1.
If you want it to be more stable,
which I want, I'm going to go with E1.
And if you want to make mobile apps,
which this platform can do in React Native,
you can use the mobile agent, but
we don't need mobile for this particular application.
Then we're going to go and change the model
if we want.
So I'm just going to use cloud Force on it.
We could change GPT five or use the extended
thinking model or reasoning model from Claude 4.0.
But as you can see, that's going to be two times
more expensive.
It's not really needed right now,
so I'm not going to use it for a few other features
here, like adding MCP servers and stuff and changing
the templates and the budget for the application.
You can also change the,
if you want this to be private, public, etc..
So I'm going to close that.
I'm just going to go ahead and press on run.
And this is going to get going for me okay.
So while it does that, let me quickly
talk about the application that I want to build
and some of the components of the prompt
that I put together.
So some of you may know, I run a mentorship program
called Dev Watch Now for this application.
I figured,
let me build something that I actually know.
You know what I want to build
that I have a lot of experience with, you know,
something that would actually solve a problem for me.
And that is kind of a dedicated platform
for all of our dev launch students.
So they can view, for example, their roadmap,
or they could see the calls that are coming up.
They can see all of the courses that we have,
something where there's, you know, a lot of different
components to it, and there's actually authentication
and there's users and there's roles.
And we have coaches and we have admin.
So we need to be able to create data
and save data and, you know, store specific data
for particular users.
We need to be able to onboard people and have them,
you know, fill in a form book calls with us,
all of that kind of stuff.
So I'm not going to go over everything.
But essentially that's what I'm asking this to do.
I give it the context, like this is the program,
this is what I want you to do, right?
Kind of my motivation.
And then I go through and I have different features.
We need authentication and roles.
We need user management.
We need group calls. We need courses and lessons.
We need an onboarding flow.
And then I'm going to add
an AI matchmaking component.
Running time someone joins, it will automatically try
to match them to another student.
That's similar to their experience.
So they can connect, collaborate and kind of,
you know, build a community.
Anyways,
that's kind of what I'm attempting to build here.
And that would solve a big problem
for us at Dev Launch,
because right now we have all kinds of software
that we're using, and while it's fine,
it can be a little bit hard to organize
and get people onboard into that.
So this would be better.
So hopefully got some context there.
I went back to emerging now
and we can see the agent is asking me some questions.
Says, you know, before I start
can you answer these for the AI matchmaking.
What do you want to use.
So my case I'm going to use the emergent LMK.
Now the way this works in emergent
is that it has a universal look
for all of the popular kind of AI models
like anthropic, you know, open AI, etc.
Gemini, as you can see here.
So rather than having to provide your own key,
you can just use the universal key
which makes the integration very easy.
So I'm going to go ahead and say yes,
let's do that now.
I also had the AI the ability sorry for users
to like reset their password or something.
So it's asking me
what should we use for the email service.
So I'm just going to tell it to use SendGrid,
which is one of the options there.
And by the way, if you didn't understand the question
you could just ask the agent, hey, can you
clarify this? Can you tell me what you mean?
I don't know what SendGrid is,
what do you suggest, etc.
for authentication?
How should the first admin account be created?
I'm going to say create it for me.
Use you know Tim at SAC with Tim
dot net and admin 123 as password.
Okay.
And then for the Qualcomm integration
do you have a embed code ready.
Should I create a placeholder.
Use a placeholder for now because I told it
hey, we're going to have to have like some bookings,
you know, for one on one meetings and stuff.
And we're going to use Qualcomm okay.
So I'm going to answer all of those.
Press enter
and then it should go ahead and start working.
I will warn you this can take a long time.
Can take 15 20 minutes.
Sometimes depends on how much you're asking it to do.
And usually the first prompt takes the longest.
And then the ones after that are going to be
a little bit more iterative.
Now, while this is building,
I do want to explain to you a few things
about kind of the architecture of the type of apps
that you're likely going to build.
Now, when we build an app and we build a mobile app
or a web app or something like that,
you have kind of three main components
you should be aware of.
The first is the front end, then
you have the back end, then you have the database.
Now the front end is
kind of what you're seeing right now.
It's a user interface.
It's the thing a user is going to look at.
It's where data is presented.
It's where someone maybe signs in, for example,
where they make a post.
Right.
So the thing we're all familiar with now,
the front end is typically pretty simple to build.
And that's why you see a lot of other popular
AI tools that just generate front ends for you.
Right? Or make really nice landing pages.
Now here we're trying to make something
a little bit more complex.
We're going to have user
data. We're going to have rules.
We're going to have permissions.
We're going to have kind of this like content
management system where we're creating calls.
And in order to do that you need to have a back it.
Now the back end is where secure operations
for your application can take place.
For example, authenticating a user
and making sure this users who they say
they are checking their username and password. Right.
Associating different data
with different users, having roles
like I'm an admin
and then other people are students, for example.
That's really what the back end is responsible for.
The heavy logic
and anything secure for your application,
anything that you wouldn't want
any random user to be able to see?
Well, you need to kind of gate that
by creating a backend, right?
And that mostly has to deal with authentication and
authorization, which emerge and can handle for us.
Then we have the database. Pretty straightforward.
But that's where we store all of the data
that's usually connected to the back end.
And the flow is that we have a front end user
does something that sends a request
to our back end back
and make sure the user can actually do that thing,
and then it does the thing essentially pull
some data, return some result to the front end.
Okay.
So just understand front end back end database
and the reason why this platform
kind of stands out to me and why I'm using it for
this tutorial, is that it can handle all of that.
So you can create a full stack application,
not just a front end
and something that is actually quite complex,
which is what I want to show you how to do.
So it's been a few minutes
now, and the agent has just completed.
And what I'm going to do now is take a preview
of the application and see if it's working.
Now, that's
usually the way that I would do things right.
Ask it for something, check
to make sure it actually did the thing you asked for,
if it didn't revise
or tell it to fix it and go on from there.
Now, the way that emergent works,
which is a little bit different
from some of the other tools,
is it actually automatically tests things for you.
So there's multiple agents
that are running in the background.
One of those agents is an automated testing agent,
where it's
going to make sure that everything you asked it to do
is working properly.
For example, during this build,
there was an error with one of the files.
It detected that error, it automatically fixed it.
And you can see if you look through
kind of all of the logs that are appearing here, it
says, you know,
I noticed the login screen isn't working.
Okay. Let's try to fix that okay. Perfect.
The test results are showing
the application is working right.
Because it has that agent
that's doing the testing for you in the background.
But of course if you run into a bug,
well then you can, you know, just tell it to fix it
or you can even dive in there and fix it manually.
So what I'm going to do is I'm just going to refresh
here the app preview window.
You can see us as welcome to Dev Launch sign
in to access your mentorship program.
It told me that this was the account that it created
and that's what I asked for as well.
Admin 123.
So let's go ahead and sign up okay. Cool.
So it brings us into the admin panel here.
Keep in mind this is all from a single prompt.
There's a few buttons that aren't working
like view schedule you know find my study partner.
If I go to this
looks like that page doesn't work or it crashes.
Maybe if we go to group calls again,
the redirect doesn't seem to be working properly.
So that's a bug that obviously
we'll need to fix. Right.
And then we have search.
That also doesn't seem to really work properly,
but generally it looks like
the application is kind of functioning,
but there are definitely some bugs.
So let me share that now with the agent.
Tell it, hey, you know, these things aren't working.
Let's fix it.
And this is, by the way,
a huge part of vibe coding, right?
You're going to ask for something. It's
going to mess something up.
There's got to be a mistake.
There's going to be a bug.
Now, these tools are getting better and better,
but really this is a test of patience to see, okay.
You know, are you actually able to sit there
and go through all of these problems and fix them,
which now is significantly easier with the AI models.
Whereas before, as a developer, you'd have to
wrack your brain and go fix them manually yourself.
Anyways,
I'm going to just kind of write a prompt here,
tell it what's going wrong,
see if it can fix that, and then we'll go from there.
We'll talk about some more features. Okay.
So I've just written the prompt here again
I just told it
I can't really interact with anything on the site.
Let's go ahead and press enter
and let's see if I can fix that. All right.
So it looks like it. Just fix this.
It told me hey, you know I did this stuff.
Can you make sure that this is working now.
So let's refresh the live preview here
and let's see if we can navigate in.
Perfect. There we go okay.
So now I can navigate to the different areas.
I don't know if all this stuff
is going to work right now okay.
So these buttons still aren't working okay.
Completed onboarding
join dev launch user management doesn't look
like I can click into the users, but overall
a lot of the functionality is here.
Okay add user. Let's try Tim at
gmail.com.
Tim. Are okay creates.
All right.
Looks like that was created okay. Cool.
User created invitation.
Email has been sent I'm not able to test.
That's not a valid email but that is a good start.
All right.
So this is pretty good for a start.
Now I'm just going to tell it.
Yes all of this is good.
But before I do that
I'm going to save this to GitHub.
So if you've never used version control
before there's this tool called git and GitHub.
GitHub is kind of like the online version
where you can create what's called a repository,
which stores all of the code that's been written.
Now behind the scenes
here, emergent has wrote a ton of code.
And what we want to do
is make sure that we save that code
so that if we need to go in and manually fix it
later on, or if there's a bug that pops up later,
we can get back to a previous version
now built into the emergent platform.
You can actually roll back.
So you see, there's this little thing.
This is roll back here.
You can press this button and it will bring you back
to the previous state and change
all of the code to be at that position.
However, git and GitHub is more universal.
It means we can also, for example, take this code
and use it on our own computer if we wanted to.
So what I'm going to do is just press this button
that says Save to GitHub,
and I'm going to connect my GitHub account to this.
Don't worry, GitHub is completely free.
Just make an account on GitHub and then connect it.
And if you're using any type of vibe coding tool,
I highly recommend that you use this immediately.
And if you don't have some kind of, GitHub,
you know, integration
built in, then tell the agent, hey,
I want you to make a git repo.
I want you to explain to me
how to upload this to GitHub
and ask it explicitly how to use git.
I'm not going to go over all of the commands
and everything for git, but
generally you always want to be using this,
especially in a larger project,
because if there is a nightmare situation,
you'll be able to look up like
how to get back to the previous state,
assuming that all of this was tracked on GitHub.
Okay, so let's go ahead and connect to GitHub
and let's go through this.
I just need to authorize this for my,
what do you call it kind of organization.
I'm going to install an authorize.
And then obviously I put in the code.
So let's do that.
All right. So I just connected it here.
Now I will do this in one second.
But for now I'm just going to press save to GitHub.
And it's going to ask me for a repo.
What I'm going to do is go and press on create new.
And for the repo I'm going to enter a name.
So I'm going to call this dev launch
and we'll call Student App or something like that.
And for the branch just leave this as main.
And I'm going to go create and push that to GitHub.
All right.
So it looks like that was successful.
I'm just going to press view on GitHub
and just open this up.
And you can see now that we have
all of the different code that was written right.
We have the back end, we have the front end right.
All of the different components C etc.
all of this kind of stuff.
And we are kind of good to go there
from the GitHub side.
So again, you could
then take this to your own computer.
You could bring it into another tool.
That's why we definitely want to be using GitHub.
All right.
So let's go okay.
Got it I'm going to tell this it is all good.
And then we're going to continue
and get it to continue building
the rest of the features of our application
and make this complete.
Okay. So it's now waiting for my reply.
Again.
It's pretty much saying okay, you know,
I've done all of these kinds of things.
This is the status of the application.
But we know that there's more
that we want for this app, and things
obviously aren't quite working yet
or we don't have everything 100%.
So I'm just going to tell the agent that
before we do that, though, I'll quickly
go over a few features you should be aware of.
So first of all, you'll notice that I'm kind of
in this new view because I closed the live preview.
So if you close that here, right.
And you can see now that we have this code
editor view.
So if I click on code, this is going to allow me
to actually open up a VS code.
Kind of code environment in my browser here
where I could make changes directly to my code base.
And those will be automatically
synced here with emergent.
So it says, okay, this is the password.
I did not save the password that I should have saved.
So let's go code and copy the password
and go back here and paste that and submit.
And that should bring us
into this coding environment.
Now I'm not going to write any code,
but I just want to show you that this does exist,
because if you are a developer
or you want to change a color or an image or
something like that, then you can do it from here
and you have a live coding environment
directly in your browser
where you can view all of the files.
For example, we go see, we go components, right?
We can see we have a ton of different components
that were created like label, menubar,
you know, pop over all of this kind of stuff, right.
Same thing.
If we go to the back end, you'll notice
that this isn't using like super base or some other,
you know, kind of plugged in back end.
It's actually a custom backend here
written in Python.
Now this file is quite long, so we probably
would want to make this a little bit cleaner.
But you can see all of this stuff that's actually
happening, especially if that means something to you.
Okay. So that's the coding component.
I would say generally if you're vibe coding
and you don't know how to develop,
probably don't dive into the code too much.
You can read it, you can search for things,
you can change small values.
But generally trying to make large changes
is probably going to cost
you more than it's worth
if you don't know how to code.
If you do already know how to code,
of course you know what you can and cannot do there.
Okay, so Adrian is waiting.
Let's give this another prompt now and kind of
tell it to keep going with the rest of our features.
So I'm starting small here
and I've just given it a prompt telling it, hey,
what I want to do is I want to work on the course
module feature right now
rather than telling it like build everything
and having it work on, you know,
20 different things at once,
which these agents are capable of doing.
I want to go one feature at a time.
That way I can build a feature, I can test it.
Once it works, we can move on to the next one.
And then I have really nice check points
where if I do need to roll back
or I do need to, for example,
go back to a previous version.
I didn't build this massive amount of code,
break everything,
and then really reset my progress drastically.
So I'm going to go ahead and press enter.
We're going to build that component.
Then we'll move on to the next ones, and
we'll keep building out this app until it's complete.
All right. So it just finished.
It's telling me now that it has completed this kind
of course management system.
So let's open up the preview
and let's test it out and see if this is working.
Okay. Let's go courses and cool okay.
So we have complete test course.
If we press on start learning we can get
to the modules which is what I was looking for.
We can press watch and then we can watch a video.
I just told it give me the video on YouTube.
So that's why it's linking to like a random
YouTube video for right now
says the duration has a description.
And that's pretty much what I was looking for
in terms of the courses.
Now we could have some images and thumbnails
and stuff, but I think for now that's okay.
And then I want to test
if this works. I'm going to say add module
new module
okay testing and the order I don't really like that.
But again it's okay for it now okay.
Module two at a sub module intro
create add a lesson lesson to that da.
The okay.
And actually that's not going to work https
youtube.com.
Duration ten create lesson.
And it looks like that is working okay cool.
So I can see that it's functioning here.
We can go into the various courses here.
We can see all of the lessons and whatnot.
And now what I want to move on to is the group calls.
So effectively
let's do the same thing for group calls.
Then we can do the AI component.
We can just keep building feature by feature.
And notice again
because we kind of took this small feature.
It takes less time obviously for this to run.
We can test it really easily.
And it's a lot more clear
what the app actually needs to do or what
the AI needs to do, rather than asking it
to build a massive thing at once,
which oftentimes is going to lead you
to a lot of bugs and a lot of mistakes.
Okay, so I just gave it a long prompt here
on effectively saying, hey, do the group
call feature now.
I gave it a bunch of information
on what it needs to do for that.
And we're going to go ahead and press on enter.
And now we'll simply wait for that to finish.
All right.
So this one just finished here again a few minutes.
It said okay essentially build
the thing that I asked it to build
but tells me that well, it
automatically tested itself and it looks promising.
I should test it manually,
which you should always do again.
Well, these tools don't have the built in testing.
In this case,
there's an agent in the background testing for you.
So let's go to group calls okay.
And we can see we have a few calls.
It looks like I can edit okay. Very cool.
I didn't even ask it for that.
But it doesn't look like I can delete.
And that's because it was ended.
Let's see. Yeah.
There's no way to delete.
So I'll quickly ask it to add that.
However, looks like join call.
Yeah just redirects me to zoom, which is fine.
And then can I make a new call?
Let's test this.
Let's call, you know, dev with Sam or something.
Select the host.
Okay.
I'm the only admin, so it's me.
Description will be a group call.
Okay.
Let's go. Date time.
Just make it like this.
Maybe meeting link.
Okay.
Soon.com.
And then schedule okay cool.
And then we have double team
showing up here in one day.
Nice.
So I think the only thing I need
is the ability to delete the calls.
Let me ask for that.
And then while we do that
I will add another feature as well.
All right.
So I just went through the flow of actually
getting it to the delete features.
You can see now that I'm able to delete stuff
same thing in the courses
like there's a delete icon now which I asked it for.
And then I told it
hey I want to test the onboarding flow now
because I don't know what it looks like
to be a normal user.
I just know what it looks like to be an admin.
So can you create like a new normal user account
for me so that I can sign in with that user and test
it? That's I mean, that's kind of what I asked it in
like a less verbose, you know, wording.
So let's see if it creates a new account for me.
And then I can sign in, test
that and see if it's working.
And again, the cool thing here
is that this is a native built in back end.
So rather than using super base
or having to deploy this to even test it out or run
this stuff yourself, it's just automatically running
inside of a machine by default.
It's handling all of the creation.
You can see that beep that you guys
probably just heard means that it's completed.
And now it's telling me, okay, sign in with this
with test pass one two, three.
So let's sign out okay.
And let's go test pass 123. Sign in.
And did I spell that right.
Test pass.
Let's see 123 test user at dev launch.com.
And it says login failed okay.
Well that's a good bug to find.
Let's tell it that that failed okay.
So it's fixed that. Now I was able to sign in.
It says enter your full name.
So let's go Timothy Seeker profile photo okay.
We don't need one right now.
Tell us about yourself.
What is your current programing experience. Advanced.
What technologies are
you most interested in learning?
Okay, looks like there's no box for that.
So again we could fix that.
What are your main learning goals?
I want to get better at
interviewing.
How many hours per week can you dedicate to learning?
Let's go to 30 plus and go next. Okay.
Almost done. Book your strategy. Cool.
And then this is where
the calendar integration would be.
I don't want to expose that link right now
because obviously that's for my students.
But this looks cool.
And then I can go complete onboarding.
And okay another quick bug.
So it says onboarding completed.
But then it brings us back to this page.
So let me tell it about that.
And essentially describe what happened.
And then it should fix it.
Now again
what's happening right now is going to happen right?
When you build a large application,
you're going to have bugs.
You're going to need to test it.
Just make sure you're going slowly
and then explain really
clearly how to reproduce the bug
and what the error was.
This is going to help the AI solve it.
So don't just say something's wrong.
Tell it exactly what is wrong.
What did you do to get to that result?
What was the expected outcome
and what happened instead?
This is kind of a standard like bug reporting
framework, but what we're going to do is share.
How do we reproduce it, what was the error
and what were we expecting to happen.
All right. So describe the error here.
Just so you know what I would have said here
right after I completed the last onboarding step,
redirect me back to the beginning of the onboarding
where it said, complete your profile.
I was on step three.
I pressed continue
it said onboarding completed successfully.
We saw a little message pop up. Right.
But then it didn't redirect me to the correct page.
It brought me to the beginning of the onboarding.
So that's pretty clear in terms of what happened,
what the error was.
That will help the I get to the problem faster.
So let's go ahead and press enter
and see if it fixes that for us.
All right.
So so this is fixed here. So let's go back.
Same thing. And let's test this.
Although I guess I'm going to need to refresh
and let's see.
And there we go okay so it brings me to the
correct page because now the onboarding is completed
and looks like I don't have the ability to delete
because I'm not a admin.
I can see the dashboard,
I can see the group call. Same thing.
I can't edit them.
Okay perfect. So that's fine.
That's working good. Sorry.
And then if I go to find my study partner it says no
matches found says complete your onboarding.
Check back later for study partners okay.
So I think that's the next feature
that I want to work on.
So let me tell the
AI to start building that that is now working.
However the find your study part
your partner feature is not working.
Can you create some fake data in the database
so that the find
your study partner
feature will work with the AI Lam integration?
And by the way, when I press on find my Match,
it says that my onboarding isn't completed.
Even though I'm inside of the student dashboard.
Let's make sure that all of the data is correct here
and that this feature is working.
Okay, so I told it,
you know, fix the AI matchmaking feature.
I just ran it right now
and you can see that it's working.
It says, you know, both have advanced
experience levels, found the match percentage
then gave me a few results.
This connect button isn't doing anything right now
because I didn't specify what I want that to do,
but it just shows a few students
that you could like reach out to.
And then it has kind of some boilerplate,
you know, fake images here, etc..
Cool. Okay, so that is working.
I would say overall
most of the features that I want are built.
Again, some of these things, like the navigation
is incorrect, which we could fix later on,
but I'm not going to drag you
through that in the video.
Also, this doesn't make a ton of sense
in terms of collapsing the nav bar,
but generally speaking, most of it is built.
So what I want to do at
this point is go over a few more advanced features
that you want to be aware of
for these types of tools.
Now the first is this fork mode.
Okay.
So let's explain kind of what this does
every time that you use this LLM, right.
And you use this AI tool, it's preserving
all of the context that you had previously.
So for example,
this really long conversation chain, right now.
Right.
And there's all these different things
that we've said.
And the AI is tracking all of that.
And it keeps essentially building
longer and longer prompts
with more and more context, which are more expensive
and end up costing more credits.
So what we can do is we can fork the current session
that we're on,
and what that's going to do
is take all of the context, actually
turn it into a really smart summary
that's significantly smaller and less tokens,
which means we'll end up using less credits
and then pass that to a new session
so that we don't have all of this information
that we probably don't need.
So when you're working on larger
sessions, this becomes more important.
And what we can do is preserve the context, right.
Refresh the memory and then spend less.
So especially if we just want
to start from a clean slate.
We just want to do something
that's kind of a separate like independent task,
but we still want the previous history.
Then we fork this.
So I'm going to go to fork.
It's going to generate the summary
of all of the messages that we had here.
Then it's going to take that summary pass
that to kind of our new session.
And we'll start fresh from there.
And we can continue working on this adding
new features etc..
All right.
So just made this summary for us.
Obviously we can read through it
if we want in terms of product requirements,
key technical concepts, code architecture,
all the stuff that it has done.
A lot of this will not be extremely relevant to you,
but you can see that it's done
that if we wanted to, we could edit it.
In my case, I'm going to approve and continue.
And then what I'm going to do is fork
the current agent as well.
So I'm just going to use the default agent.
Now we could change the model that we want to use.
And there also is some more advanced features here
to create your own agents that are running
running in dedicated VM.
Sorry, but I'm not going to dive into that because
it's not super relevant to this type of video.
So anyways, I have forked the new session.
It's going to do this.
Once that is done, I'll be back and we'll continue.
Okay, cool. So it's promised this new session here.
You can also see if we click through here
that now we have a new window up top.
Because we forked over this session
we could continue working in this one.
We can also work in the fork session obviously,
which is going to be cheaper now.
And then we can preview the application,
look at the code, all of that kind of stuff.
I'm not going to keep building anything else,
but I will show you a few more things.
So you should be aware of.
So let's say that we're finished with this now,
right?
All is done.
What we can do is we can deploy this now.
We can literally just press the deployment button.
When we do that, it's going to actually host
the application, not in a preview environment,
but in a production environment
that people could start using.
Now you can see it cost 50 credits per month.
So decide obviously if you want to do that or not.
And then if you want,
you can just save the current app to GitHub.
Right. And you can go deploy it yourself.
Now that's going to be a little bit complex
unless you know how to code.
So probably a lot easier and honestly
even cost effective just to deploy it in this.
But if you wanted to deploy it on, you know,
your own server or something, you can do that.
Because if we go back to GitHub, right,
let's refresh, you'll see that
we have all of the commits and all of the history
of like everything that we've done here in GitHub.
So I'm going to go ahead and press on deploy.
Now since it'll take 5 to 7 minutes.
So we'll wait for that deployment to finish.
And then we'll check out the completed app.
So a few minutes later and the app is now deployed
we can visit the site also.
Let's go back.
We also could have connected our own domain.
I'm not going to do that at this point.
But you can
and you can see it brings us to our live page.
Now let me try to sign in okay.
And I successfully signed in.
And you can see that I can go through and everything
that we saw before is functioning here.
You know, I can edit a call, I can what is it?
Delete something.
You get the idea.
Okay. So the deployment is finished.
And that's pretty much everything that I wanted to go
over.
Now in terms of how many credits this cost us,
you can see
per kind of, I guess, chat or session.
So in this session, which is relatively small,
it just cost us 2.28 credits.
The way you saw that is if you close the tab here,
so you press the X button,
you can press on this little info button.
And on the info button it shows you the credit
spend, the machine type, the model job ID,
all of that kind of stuff,
and also the most recent export to GitHub.
And then if we go back to the other one here
and we open this up,
you can see that cost us about 55 credits.
That was a lot obviously,
because it did a ton of stuff.
And then same thing.
You can scroll through here and see everything else.
So just something worth noting.
You can check the credits set.
Now last few things to mention here.
If we go back to home,
there are a few other features.
That you can use here. For example, we can go to MCP.
We can add our own custom MCP servers
here, like Super Base
or Notion or something along those lines,
or our own custom MCP.
We also can change the template,
as I mentioned before, change the budget, etc.
so if I go here and I go to new workspace, if I want,
I can create a shared workspace with multiple
other people so they can collaborate with me
and we can work on the same project again.
I don't have a team right now, so
I'm not going to do that, but something worth noting.
Okay, so that mostly covers
everything that I wanted to go over here.
Now, I know these videos are a little bit
all over the place.
That's just kind of the nature of vibe.
Coding is unpredictable. I didn't script this.
I wasn't sure exactly
what was going to happen in this video, but generally
what being successful with these tools comes down to
is how well do you plan?
So how well do you actually know what you want
and how kind of deeply have you thought through that?
How do you craft your prompts and instruct the model?
And then really, just how persistent are you,
and how often are you willing
to deal with bugs, errors,
and fix those and get through it?
Programing and vibe coding is ultimately just problem
solving.
These models obviously do a lot of it for us,
but the better we are at problem solving
and more importantly,
kind of breaking down our thought process
and planning out what we want to do
into small logical tasks.
The more successful you're going to be here anyways.
Guys, that's going to wrap up this video
if you enjoyed, make sure leave a like subscribe
and I will see you in the next one.

Video Length: 34:46
Uploaded By: Tech With Tim
Published: 10/6/2025
View Count: 20,917

Related Software Products
Emergent
Emergent

Published By:
Emergent Labs

Description:
Emergent: Build a Web or Mobile App by Describing It in Plain English Emergent is a browser-based AI app builder: you describe the app you want in a chat, and its AI agents write the front end, back end and database, test it and publish it at a live web address. It's for founders, small businesses and non-programmers who want a working prototype or internal tool without hiring developers, and it costs $20 a month (or $17 a month billed yearly)...


Related Videos
This Coding Tool Is For Real Business People; Emergent Review
This Coding Tool Is For Real Business People; Emergent Review

Video Length: 09:30
Uploaded By: Shark Numbers (10/17/2025)
View Count: 1,048,922

Emergent AI Review (2026) | Don't Waste Your Money!
Emergent AI Review (2026) | Don't Waste Your Money!

Video Length: 10:43
Uploaded By: Software Scope (3/5/2026)
View Count: 22,356

Emergent Review 2026 - My Honest Opinion (Not Sponsored)
Emergent Review 2026 - My Honest Opinion (Not Sponsored)

Video Length: 13:36
Uploaded By: Venture Harbour (1/7/2026)
View Count: 7,117

Emergent Labs: Build production-ready apps through conversation
Emergent Labs: Build production-ready apps through conversation

Video Length: 10:02
Uploaded By: Developers Digest (10/15/2025)
View Count: 1,696

Copyright © 2026, Ivertech. All rights reserved.