AI Summary
The creator provides an update on a project to build and sell iOS apps using AI automation, reporting $275 in sales from three apps over two weeks. He demonstrates a new AI-powered app called 'Looks' that uses GPT-2 image API for styling recommendations, and discusses plans to scale by trend-surfing viral app ideas.
Chapters
Total sales of $275 from three apps: Nido Collector (94 sales), Poke Machine (26 sales), and Looks (3 downloads). One in-app purchase on the AI app.
Cloud Code automates the entire process: opening Xcode, launching simulator, writing code, and testing apps autonomously.
The app offers beginner sessions, photo upload, and single-session in-app purchases. It uses GPT-2 image API to generate old-money lifestyle images, haircuts, and wardrobe palettes.
GPT-2 image API costs: $8 per million input tokens, $30 per million output tokens. Developers must calculate per-session costs to avoid losing money.
Next app will focus on trend-surfing viral ideas like 'Umogle' (looksmaxing trend) to minimize marketing effort.
The project is profitable with minimal effort, and the creator plans to scale by rapidly building apps around trending topics using AI automation.
Clickbait Check
85% Legit"Title accurately promises an update on AI app earnings and automation, delivering on both."
Mentioned in this Video
Study Flashcards (8)
How much total revenue did the creator report from the three apps?
easy
Click to reveal answer
How much total revenue did the creator report from the three apps?
$275
Which app sold the most units?
easy
Click to reveal answer
Which app sold the most units?
Nido Collector with 94 sales.
What tool does the creator use to automate Xcode and simulator tasks?
medium
Click to reveal answer
What tool does the creator use to automate Xcode and simulator tasks?
Cloud Code.
01:30
What API does the Looks app use to generate images?
medium
Click to reveal answer
What API does the Looks app use to generate images?
GPT-2 image API.
04:30
What is the cost per million input tokens for GPT-2 image API?
hard
Click to reveal answer
What is the cost per million input tokens for GPT-2 image API?
$8 per million input tokens.
04:30
What is the cost per million output tokens for GPT-2 image API?
hard
Click to reveal answer
What is the cost per million output tokens for GPT-2 image API?
$30 per million output tokens.
04:30
What monetization model does the Looks app use?
medium
Click to reveal answer
What monetization model does the Looks app use?
Single-session in-app purchases.
02:45
What is the creator's strategy for choosing new app ideas?
medium
Click to reveal answer
What is the creator's strategy for choosing new app ideas?
Trend-surfing: picking viral trends with search demand and market gaps.
06:00
π‘ Key Takeaways
$275 Revenue Milestone
Demonstrates real profitability from AI-automated app development with minimal effort.
Cloud Code Automation
Shows a practical tool for automating iOS development tasks, increasing efficiency.
01:30API Cost Calculation
Highlights the critical need to calculate per-session costs to avoid losses when using paid APIs.
04:30Trend-Surfing Strategy
Reveals a scalable approach to app development by capitalizing on viral trends without marketing.
06:00Full Transcript
Okay, so about 10 14 days ago I made this video how I made money automating iOS apps with three with AI in three days. And I thought we'd do an update now, so we're going to check how much money we have made so far and what my plans is going forward with this. And we're going to take a look at AI app test I just did basically how I set that up. So, yeah, let's just get
into it. And if we go to our App Store Connect, let's just take a look at the apps we have now. So, basically I have only done like three apps now that can make us money. So, the latest one is my testing app using actually AI to build an app. It's called Looks. Yeah, you will see it later. But Nido Collector is the one we created for sale and the Poke Machine app. So, these two are
kind of our 13 14 days running apps. This is just a few days ago since I made this. >> [snorts] >> And these are just test apps I did. This I did for the previous video. But let's head into the sales and trends now and let's see how much we have made. So, if we zoom in here, drumroll please, $275 so far. That is very good. I'm super happy with that. And if you kind of look
at what where those sales come from, you can see our graph here. Pretty steady actually. So, I'm very happy with that. And most of the sales have come from the Nido Collector app. We sold 94 of that. The Poke Machine has sold 26 and we have three downloads for the Looks app. Yeah, that's not I just put it up. So, totally we have sold 132 or maybe around 130 or something like that. And we have one
in-app purchase on the AI app, right? So, so far this is working pretty good. I'm super happy how it turned out. So, today I just wanted to show you kind of the latest app. I want to talk a bit about how we're going to scale this and how I plan to follow up this project on this channel. So, yeah, 275. That's just great for the effort we put in. So, if you didn't watch my previous video,
remember we did this everything automating. So, we automated basically the full process. So, you can see I'm here now in Cloud Code and I can just say open Xcode with the app Looks in Xcode simulator, something like that. And now we should be able to just open this by using Cloud Code and we can actually edit the app from there. So, you can see we are launching Xcode down here now and you can see we go
straight into the app. Perfect. And let's see if we can open up the simulator too. And here you can see we launched the app here now. So, you can see Cloud Code can do all of this for us. It even Of course, it writes the app too, but it can also launch the simulator and it can test the apps for us autonomously. And you can see here we have it. So, this is the Looks app, right?
You can see we can do like beginner session and you can see we can scroll down here. This is where It's basically a simple app, right? You can put in your We can do beginner session. You can open like a photo here. So, let me try to do that. So, we have a test photo here, right? We can confirm this. We get the privacy and here is kind of the payment structure. I think I'm going to
show you this on mobile because it's a bit better. So, you can actually see how we actually choose to monetize or take payments for this app. Okay, so you can see we are here kind of on the same, but this is a bit smoother, right? Because we are inside the app. So, I'm just going to click beginner session and I'm going to upload an image here. So, you can see we just grabbed this It's another AI
generated image. So, let's say we have her. So, we're just going to center, confirm. And here we can pick like a session. So, we can do whatever we want here. We can do like one session and what is nice now is if if I select one session, you can see we get this here. So, we actually do single app purchases. So, you pay for only what you want, right? So, if you want five sessions, you can
use this five times, right? So, that is kind of the way I choose to monetize this. And yeah, I guess I could show you a couple of results too, but that is not the most important part of this video. It's basically how we choose these selections. And of course, this will run the OpenAI GPT-2 image API. So, of course, we got to cover our costs. So, this is in Norwegian, but it's like yeah, some kind of
dollar amount here. I can't remember now. But basically that is the setup we have and this was of course approved in the App Store. So, let me show you like some results of this. So, this is of course the not the most insane app, but basically you can see here when you upload like an image, you get your image in like four images from like the old money lifestyle, just four random images. And this could be
male or female, so it does kind of look at if you are male or female when you upload image. And here is kind of the saloon. Here you kind of get six different haircuts. This is also true for females, right? It's a bit different hard to scroll on this, but you get these old money looking hairstyles, right? And I think the final thing I had is like wardrobe palette, so you can kind of see what kind
of clothing that matches your style both for female and male. And that is basically my test app. And like I said in the video, this runs on runs on the GPT-2 image API. So, here is where you kind of need to do some research if you want to do this because here the it matters, right? The pricing matters and stuff like that because you need to calculate how much you should yeah, actually do per sessions. So,
you kind of know you can make like a small profit. I just I didn't really make this to make money. I just wanted to test out how we can actually build a pipeline to make it template-y for future versions. So, let me just quickly show you the pricing here. So, if we look at GPT-2 image, you can see input eight cached input. That is also something and output 30. But that is not per image. This is
per million tokens. So, you need to use kind of a calculator to figure out if you look at kind of like a medium quality or you let's say you do high quality GPT-2 image. And then you kind of need to look at is this going to be like vertical or horizontal or one by one. And this is kind of where it kind of comes in. If you want to build an app, you need to do some
calculations here to figure out so your app doesn't lose money, right? Because that could happen. And of course, you need to test it, test kind of the code so it doesn't don't make any mistakes and do like, let's say, 10 15 20 calls, right? Because this is quite important when you do an app like this. So, this will of course show up as an in-app purchase if you do what I did now. And yeah, that is
the best way I have found so far of actually doing these AI apps. And this is just the beginning for me because I wanted to quickly talk about um my future plans and the next AI app we're going to build, but that is going to be in our next episode, but I guess I could show you what I have in mind so far. Okay, so what I was looking at is this Umogle. This is basically like
a viral trend I've seen like on Reddit, X and stuff. And it's like a yeah, it's like an image based app I think you kind of face up with someone, some kind of looks maxing thing. I'm not quite sure. But basically this is the category I want to focus in. This is like a trend like trend surfing because I want to be able to pick like an app idea that doesn't have to be long-lived, but basically
you can be created like fast, it's trending, people are searching for it, there's a gap in the market. So, basically you don't have to do any marketing. So, I think for the next episode in this project, we're going to build out a new app and it's most likely going to be an AI app so we can see a bit of the calculation how you need to do that because making these static apps is not that advanced
when it comes to this. So, if you want to see kind of how I did this in practice, you can go back and watch this how I made money automating iOS app with AI in three days video that I did a couple of weeks ago. And yeah, basically that's what I wanted to share today, a quick update on how the sales have been and what I want to continue working on and how we're going to try
to scale this to maybe get the sales up even more. Who knows? So, yeah, you can see this has been like $20 a day or something. So, that's pretty good for that low effort using Cloud Code and stuff. So, hope this gave you some inspiration and you want to follow along this series. So, thank you for tuning in and I'll see you again soon.