Remaking Popular Apps (To Avoid Paying For Them)

Category: AI Tools

Tags: AICodingDIYSoftwareTools

Entities: Assembly AIChatGPTClaudeMicrosoft PaintWarp.devWhisper Flow

Transcript

00:00

I'm kind of cheap. I hate paying for simple apps that I feel like should have been a free Chrome extension or something.

So lately, I've been using AI to whip up my own custom tools. Nothing super fancy.

No billion dollar startup ideas. Sweet Baby Rays barbecue sauce.

00:15

Nothing I intend to like actually sell. Just quick scrappy apps that solve little problems for me.

And honestly, it's way easier than you'd think. So today, I'm going to show you how you can use AI to stop overpaying for underwhelming software.

Let's get into it. The first app I want to create is

00:32

well, not something that you normally have to pay for actually, but I don't have access to it on my Mac, and that's Microsoft Paint. Sometimes you want to just like edit an image real quick, like circle something or sketch something on it real quick, and on a Mac, you don't have that super simple software like

00:48

Microsoft Paint. So, let's go ahead and build it.

For this first one, I'm actually going to test warp.dev. It's an agentic development environment.

And if we check out their pricing, they do actually have a free plan where you can try it out. So up to 150 AI requests

01:05

per month. You can use Claude Opus 4.1 and GPT5 and pretty much all the best models.

So I think I can do it on the free plan. So I went ahead and downloaded Warp here.

And this is what it looks like. Down at the bottom, they

01:20

have like a terminal mode, an agent mode, and an auto detection mode. So, let's just leave it on auto.

And then for the models, we have all sorts of models available. Again, this is the free plan.

My favorite model for coding was actually 03 Pro, which looks like we don't have access to that here, but

01:37

Claude 4.1 Opus is also really, really good. I've had really great success coding with this one.

So, let's leave it on Claude 4.1 Opus. And let's see if I could just give it a prompt to build the thing.

Here's my prompt. Make a Microsoft Paint clone.

I should be able to draw, use a paint bucket, create

01:54

arrows, type, and have all the color options. I'd also like to be able to drag an image directly into the app and edit the image, and then save a new version of the image that's been edited.

I'm hoping I can just give it a prompt like this, and it will know what to go off and do. So, let's see what happens.

02:10

All right, I'll create a Microsoft Paint clone with all the features, etc., etc. And then it created a task list.

This is how you know it's kind of an agentic workflow is when it creates a task list and then starts working through all of the tasks. That's usually a pretty good sign that it's going to work out well.

02:27

But let's find out what we end up with off of just a single prompt inside of warp. Okay, so it looks like once it goes through each step, I do have to apply changes.

I wonder if there's like the equivalent of a yolo mode here, but let's go ahead and apply changes. and it

02:44

has moved on to the next step. Now, down in the very bottom right corner, there is a button that if I hover over it, it says auto approve all agent actions for this task.

I'm going to go ahead and turn that on and let it just run through all of the steps. Okay, so it ran for a

03:00

couple minutes here and it opened up this app inside of my browser. Let's go ahead and give it a test.

We've got our pencil tool here. Can I draw?

Yes, I can draw. Let's go ahead and close off our loop.

Let's use our bucket tool. Can I

03:16

paint bucket fill? Okay, so I don't think the paint bucket tool is working.

Nothing seems to be happening when I click anywhere. Don't know what this button Oh, that's the arrow button.

Okay, so I can create arrows. That part works.

Can I add text? This is text I

03:32

made. Yeah, that worked.

Let's see. What is this?

Is a broom. Okay, so the broom is the erase.

I've got like a line tool where I can just draw different lines. Our square tool, our circle tool, can change brush size.

Yeah, I mean

03:50

everything's kind of working except for our paint bucket tool. We don't make mistakes.

We have happy accidents. Paint bucket tool did not seem to work for whatever reason.

Let's see. Does the save image button work?

Yeah, it saved it as painting.png. Can I upload an

04:05

image? Let's open up this thumbnail here.

Yeah, I mean it opened up the thumbnail and I could draw on my thumbnail here. So everything's working except for the paint bucket.

Let's jump back to warp real quick and just say everything's working perfectly except for the paint bucket. When using the paint bucket, it should fill in the

04:21

space with a solid color. It took just a few seconds here.

So let's refresh our page and see if our paint bucket is working. So I'll select the paint bucket.

Let's go yellow. And yeah, it seems to be working.

Let's try green. Let's uh let's draw a rectangle.

I'll

04:38

draw a blue rectangle here. And then we'll put orange inside the rectangle.

Yeah. I mean, it does what Microsoft Paint does, and I did it with a couple prompts.

Pretty freaking cool. And I drew you a self-portrait.

You know how people send voice notes because it's

04:53

faster than typing? The only downside is that they just sit there as audio.

That's why I use Whisper Flow. It gives me the same speed advantage, but everything I say turns into clean formatted text in whatever app I'm using.

Chat, GPT, Notion, Cursor, you name it. It even edits out the ums and

05:10

fixes little corrections like Tuesday, wait no, Wednesday, and formats lists, emails, or code automatically. The biggest benefit is when I'm prompting AI tools, if I type something short and rushed, I get a basic response.

But when I speak, it's even easier to give more context and get way better results. And

05:28

if you're coding in tools like cursor, warp, or claude, flow actually understands your code. It handles technical terms, tags, files, and even gets variable names right.

They also recently added snippets, little voice shortcuts that expand into full text. I

05:43

can use them for things like my calendar link or a terminal command I use often. I just say it and Flow drops in the whole thing instantly.

If you want to give it a shot, you can get an extra month of Flow Pro on top of the 14-day free trial with the code Mr. Eflow.

So,

05:59

while other people are leaving context out because typing takes too long, you'll be shipping better results by just speaking with Flow. Thank you so much to Whisper Flow for sponsoring this video.

Now, let's get back to it. Another tool that I think would be very, very helpful to have is a tool that

06:14

converts any image file into a specific image file. So, for example, I can drop in a WEBP file and it'll convert it to a JPEG.

Or I can drop in an AVIF file and convert it to a JPEG. As somebody who edits a lot of videos and does a lot of

06:31

website building, having some of these other file types like WEBP and AVIF and TIFF files and things like that can get really annoying and I like to just convert them all to JPEG files before using them in a video or a place like that. Now, there are free tools out there like this free convert that I

06:48

found and it will convert one file at a time, but if you want to bulk drop in like 20 images and have it work through each image, well, you got to pay for that feature. And their lowest plan is $13 a month for a tool where you drop in images and it changes the image file.

We

07:06

can do better than that. We could build our own that's free to use whenever we want to use it and works just as well.

For this one, we'll use Claude. And I'll just use the Claude interface directly over at Claude AAI.

And then I'll also use my little whisper tool to explain

07:21

what I'm looking for. So I'll click into the prompt box here.

I want an app where I can drag and drop any image file type and it will convert that file type into a JPEG. So whether it's a webp, an AVIF

07:36

file, a TIFF file, any of those types of files, I want to be able to drop it in to this app and convert it into a JPEG. Now, also, I want to be able to bulk drag and drop a ton of files in at once

07:52

and have it work through one by one and convert each file to a JPEG so I can bulk upload a bunch of files. I want it to be built on Python so I can run it on my desktop without even needing an internet connection.

Whisper just

08:07

converted everything I just said into a giant wall of text. I'm just going to double check it real quick to make sure it got all my ideas correctly.

Everything looks good. So, let's go ahead and submit it and let Claude build this out for us.

Claude gets to work writing the code for us. It split the

08:22

chat off to the left and we can actually see the code being written in real time off to the right. So one prompt and we've got our code written over here.

Features universal format support. So WEBP, AVIF, TIFF, PNG, BMP, GIF or GIF,

08:38

HEIC, and more. Bulk processing, convert multiple files at once, etc.

So it looks like it designed it to do everything I want it to do. And here it's telling me how to install it.

So I need to install a few dependencies on my computer first.

08:53

I have a folder on my computer called code where I put the little apps that I'm building. I'll just create a new folder here and we'll call it image convert.

Now I'm going to open up my terminal. And this should work the same way whether you're on a Mac or a PC.

We'll change our directory to make sure

09:08

we're in the right place. So now I'm in this image convert directory that I just created.

Follow the instructions that it told us to do. So we'll install pillow.

I guess since I'm on a Mac, I actually have to use pip 3 install pillow instead

09:24

of just pip. That's what claude just told me at least.

And then our next step, we copy this code that it told us to install tinter dnd2. I don't know if that's how you pronounce it or not.

Paste this in. Now it's installed that.

And it wants me to save this file that

09:40

it created for us as image converter. py.

I'm going to come to this folder, go into this image convert that we just created. We'll create a new document inside this folder and we need to name it image_converter.py.

09:56

Copy our code from over here. Paste it into our text editor here.

And we will save it. And it saved it as an RTF file and not a Python file.

So I'm going to remove the RTF and we'll use py. Now if I grab the code that it tells me to use to run this file here.

Pop this back

10:14

into our terminal. And look at this image.j JPEG converter.

It's all built and we can just drag and drop files in here to convert them to JPEG. So, I have this PNG image on my computer.

Let's go ahead and drag this image in and see if

10:30

it'll convert this PNG to a JPEG. The colors are a little wonky.

I'd probably want to fix that. But, let's find out if the software actually works.

So, I'll convert to JPEG. It asked me where I want to save it.

I'll put it in my downloads folder here. Conversion complete.

One successful. And looking in

10:47

my downloads folder, we can see that it created the image and saved it as a JPEG. Here I have this file on my computer.

That's a WEBP file. Let's go ahead and toss this one in.

All right, one file ready for conversion. Convert to JPEG.

I'll just save it to downloads again. And successful.

And there we go.

11:04

It saved that exact same image as a JPEG instead of a WEBP file. One last test.

I want to throw multiple images in there and see if it will batch convert multiple. So I'll pick a few PNGs.

I've got this one here. Toss it in.

Here's another one that I have a PNG. I'll toss

11:19

this one in. Let's throw in a WEBP file as well.

I've got this is a WEBP file. So, I'll toss that in.

So, now we have three files ready for conversion. Let's convert to JPEG.

Save them all to our downloads folder. Conversion complete.

Three successful. Let me double check.

11:36

And look at this. Man with kitten JPEG.

MW Beard inverted JPEG. And this image as a JPEG.

So, it does exactly what I want it to do. Don't have to pay monthly.

And I can drop a batch of files in here. Pretty cool.

Finally, I want to

11:51

build an app that can transcribe videos for me. So, if I'm recording a video for my YouTube channel and I want the transcription before it's live on YouTube, so I can, you know, write notes about it or get my chapter titles or things like that, I can do that.

Now,

12:06

there's tools out there that will do transcriptions for you, like Descript, which charges as little as 16 bucks a month. And there's also tools like Rev that can do transcriptions that give you 45 minutes free per month, but then you

12:22

need to upgrade for 10 bucks a month if you want to keep transcribing. So, let's build a tool real quick where I could drag and drop any video in and it will transcribe it for me.

Now, the tool I'm going to make won't technically be free to use. I'm going to need to use an API key from a service that does the

12:39

transcription. But when I use my API key, it's going to allow me to transcribe video and audio files, but it's going to cost like fractions of a penny per minute to do it.

My favorite transcription API is actually from Assembly AI. They do speech to text, but

12:56

if we look at their pricing, when you first sign up, you get $50 in free credits, which gets you up to 185 hours of transcription. And then you move to like a pay as you go plan.

And then it costs 27 per hour. So you could record a

13:12

full hourong video and it will only cost you a little more than a quarter to transcribe the whole thing. So we'll go ahead and use the API from Assembly AI.

I've already created an account, so I'll go ahead and log in. You can see when I log into my Assembly AI account, I've only used about $862

13:28

of my credits and still have over $41 remaining. So, here's how we're going to build this app.

We're going to use our API key from Assembly. So, I'm going to go over to API keys here on the left, and I'm going to create a new API key, and we're going to call this demo video

13:44

because I'm going to delete this API key after this video is live because you never want to share your API keys with anybody, otherwise other people will be able to use your account. So, let's go ahead and create this and then we'll copy the API key that it created for us for this demo video.

Now, to create the

14:00

app this time, I'm just going to go ahead and use chatgpt to create it. And just for some bonus points, I'm on a free ChatgPT plan.

You can see down in the bottom left, I'm logged into my free account. So, we're going to see if we can build this app without using one of the premium accounts.

So, inside of our

14:15

prompt box, I'm going to go ahead and use Whisper Flow again here. build me an app where I can drag and drop either a video or audio file into the app and it will transcribe all of the audio from whatever I upload.

We're going to use

14:31

the Assembly AI API key for the actual transcription service. And I want this to be browserbased.

So I want to be able to drag and drop a file straight into the browser, have it transcribe it, and then in the same window give me back the

14:46

transcription. So, I basically just explained it using Whisper, but it's going to need the API key that we just created from Assembly AI to build this.

So, let's go ahead and put here is my API key for Assembly AI. And I'll paste

15:02

it in like that. Use this for the transcription service.

Because I haven't used my think longer for the day, let's go ahead and turn on think longer and get it to work a little bit harder on this one. It thought for 29 seconds and now it's writing my code for me.

I tried

15:18

running it. It gave me this error right here.

So, I'm going to just feed this error right back into chatgpt. I received the following error when trying to run this inside of the canvas in chatgpt.

So, I just gave it some instructions and now I'm going to go ahead and put the error I got in quotes

15:35

so it can see it. And uh let's see if it can fix this for us.

I ran into a couple more bugs. I had to tell chat GPT the bugs and asked it to fix the bugs a couple times, but now I think we've got a version that's working.

And again,

15:51

still on the free plan here. So, if I click on preview, we can see it loads up our app here.

I don't like the background color and text choice it used like, you know, black on a dark gray background. Not ideal.

Obviously, an easy fix. Just tell it to change the

16:06

font color. But let's go ahead and drop in an audio video file here.

In fact, this is the video of the ad read that I did earlier in the video you're watching for Whisper Flow. If I drag and drop this in, we can see it uploads my file.

16:22

You can barely read it, but you can see status uploading Whisper Flow and the percentage going up. Now, we have status transcribing.

This may take a minute. And we have our transcript from the ad.

You know how people send voice notes because it's faster than typing. The only downside is that they just sit there as audio.

So, it pulled in the

16:39

whole video I just recorded and transcribed it. And if I jump back over to Assembly AI over here, click on my cost here, you can see that it cost me slightly more than three pennies.

Tools like Descript and Rev are great, but if

16:55

I can transcribe something for 3 cents versus 10 bucks a month, might be a little bit preferable. So, there you have it.

There's a few tools that I found paid versions for and went, let's just go ahead and create this ourselves so we don't have to pay a monthly fee

17:11

for it. And this is one of the things that I love using AI for lately.

It's not going to like recreate an entire like HubSpot CRM for you right now. But little apps that you might find helpful or finding little bottlenecks in your

17:26

business where you can build a mini app to solve that bottleneck, AI is amazing for that. And I just gave you three demos with three different tools.

We used Warp, we used Claude, we used ChatGpt. All of them were able to build these little basic apps that solve small

17:42

problems for us. AI is great at that.

And hopefully this got your wheels turning about other little apps that maybe you might want to create for yourself. Something that you're sick of paying, you know, $9.99 a month for because it can be built really quickly

17:58

and then used for free in perpetuity after that. and I can't wait to see what you create.

All of the apps that I've created are actually available on my GitHub. So, if you want to download them for free and build off them and iterate off them, they are available on GitHub.

I will link up my GitHub in the

18:13

description below if you want to go check them out and download them. They'll be totally available for free, open source, you can do what you want with them.

If you do build something cool, tag me on X or Instagram or Facebook or whatever your platform of choice is. Let me know what you built.

I want to see it. Maybe I'll even do a

18:28

future video where we round up some of the cool tools that you guys have built based on trying some of this stuff. But that's what I got for you today.

Thank you so much for tuning in. I really, really appreciate you hanging out, nerding out with me.

Some of this stuff is so much fun. I just have an absolute blast playing with these AI tools and

18:44

building stuff and putting things into the world that didn't really exist before I worked with AI to make them exist. So, super fun stuff.

I'm so happy and appreciative that I have people that want to nerd out and watch this stuff with me. And uh yeah, if you want more

19:00

stuff like this, let me know in the comments and I will do my best to accommodate. And again, if you like this, like, subscribe, do all the things that all the other YouTubers tell you to do.

It really helps the channel grow. Really appreciate you.

I'm done rambling. Hopefully, I'll see you in the next one.

Bye-bye. Thank you so much for

19:15

nerding out with me today. If you like videos like this, make sure to give it a thumbs up and subscribe to this channel.

I'll make sure more videos like this show up in your YouTube feed. And if you haven't already, check out futuretools.io where I share all the coolest AI tools and all the latest AI news.

And there's an awesome free

19:31

newsletter. Thanks again.

Really appreciate you. See you in the next one.