Big Idea

The argument in one line.

Because current Claude models can read an iPhone screen reliably, a Mac mini and a few real phones can replace paid automation tools and run social workflows at scale without breaking when the app UI changes.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…

An agency owner or freelancer who sells automation and wants a local, no-subscription way to run Instagram, TikTok or iMessage workflows for clients.
A builder comfortable with Cursor and GitHub who wants to see how Claude can control a physical phone through Xcode, WebDriverAgent and Appium.
Someone paying monthly for lead scraping, enrichment and outreach tools who wants to know which of those workflows a phone can replace.

SKIP IF…

You don't own or want to buy a Mac, since the whole setup depends on macOS and Xcode.
You want a no-code, click-through tool with support, because this is a code base you clone and prompt your way through.
You need compliance-safe outreach, since the video leans on phones looking like real users to avoid platform detection.

TL;DR

The full version, fast.

Newer Claude models are good enough at computer use that an agent can run a real iPhone by reading screenshots. The setup is a Mac mini, iPhone 11 or newer in developer mode, Xcode, and the open-source Kev's iOS Agents repo, which uses Apple's testing route through WebDriverAgent and Appium to tap and type. Workflows come in two kinds: static ones that replay known touch points for free, and dynamic ones where a cheap model like Haiku 5.5 loops on screenshots and self-corrects. Schedule them with macOS launch agents, and sell the result by reverse engineering clients' paid tools into phone workflows.

Free for members

Chat with this breakdown — free.

Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.

Create a free account →

Chapters

Where the time goes.

00:00 – 01:46

01 · Cold open and agenda

Pitch for the open-source farm, plus the six-chapter plan.

01:46 – 05:30

02 · The tool stack

Mac mini, iPhone 11+, Xcode, Apple Developer Program, USB hub, database, Cursor, GitHub and Claude.

05:31 – 07:57

03 · How it works

Developer testing route, WebDriverAgent, Appium and screenshot feedback, shown with an Instagram DM example.

07:57 – 12:45

04 · Building the farm step by step

Clone the repo, launch the UI, register the iPhone, prepare WDA and verify with Appium.

12:45 – 16:00

05 · Tour of the farm UI

Device pages, touch points, built-in workflows and the automations tab with macOS launch agents.

16:00 – 18:20

06 · Static versus dynamic workflows

Why static is free, why Haiku-driven dynamic loops unlock new use cases, and a live static TikTok warm-up.

18:20 – 22:11

07 · Building a dynamic lead scraper and scheduling it

A Haiku 5.5 workflow scrapes Instagram followers into a database, then is scheduled twice a day.

22:11 – 25:21

08 · Business use cases

iMessage marketing, posting to hundreds of fan pages, product reposting and cost-reduction positioning.

Atomic Insights

Lines worth screenshotting.

Apple blocks random software from controlling an iPhone, but the developer testing route through WebDriverAgent is the loophole an entire phone farm can be built on.
Once a workflow is a static sequence of known touch points, running it costs zero AI credits, so static should always be the North Star.
A dynamic agent that takes a screenshot after each step survives UI changes that break static click scripts.
Scraping 8,000 followers on Apify would run about $16, while a Haiku-driven phone workflow does it for pennies.
Pick one iPhone model for the whole farm so screen dimensions and touch points stay identical across devices.
A Mac mini beats a MacBook for a farm because it can stay on around the clock without battery damage.
iPhone 11 is the floor, because older models can't run the iOS and Xcode versions the automation needs.
Registering a phone is a one-time job: after setup, unplugging it only costs you a passcode entry on the XCTest prompt.
Phone-based posting is harder for platforms to flag as automated than computer-use or browser bots, because the traffic comes from a real device.
macOS ships with launch agents, so scheduled automations run locally with no cloud provider and no per-run fee.
iMessage marketing at scale usually costs about $200 a month as a service, and a farm phone with its own Apple ID can do it for free.
The agency margin lives in cost reduction: every $20 a month scraping, enrichment or outreach tool a client pays for is a workflow the farm can absorb.

Takeaway

Cheap vision models turn phones into automation servers

WHAT TO LEARN

Once an agent can read screenshots reliably, a real iPhone becomes a programmable machine, and the money is in replacing paid tools with workflows you run yourself.

Phone automation rests on Apple's developer testing route, which WebDriverAgent and Appium expose so software can tap, type and read a real device.
Buy one model of iPhone (11 or newer) and a Mac mini so every device shares screen dimensions and the farm can run around the clock.
Turn each working workflow into a static touch-point sequence, because replaying known taps costs no AI credits at any scale.
Use a fast, cheap model like Haiku as the eyes for steps that need judgment, with a screenshot check after each action to catch errors.
Map each app's screens into named touch points once, so the agent can debug failed runs by comparing screenshots against known targets.
Schedule finished workflows with macOS launch agents so they run locally with no cloud bill, and watch the first scheduled run to confirm the pipeline works.
Price client work against the monthly tools it replaces, since scraping, enrichment and outreach subscriptions are the margin you can capture.

Glossary

Terms worth knowing.

WebDriverAgent (WDA): A service that runs on a developer-mode iPhone and exposes the screen and controls over the web, so software on a connected Mac can tap, type and read the UI.
Appium: An automation framework that sits on top of WebDriverAgent and translates commands like click and type into actions on the phone.
Touch points: Named screen locations for an app, such as the Instagram profile tab, that the agent maps once and then taps by name when building a workflow.
Static workflow: A fixed sequence of taps and swipes that already works, so it can run at scale without any AI model calls.
Dynamic workflow: A looping agent workflow where a model takes a screenshot, decides the next action, and checks whether the task finished.
Launch agent: A built-in macOS service that runs jobs on a schedule, used here to trigger automations locally without a cloud host.
Developer mode iPhone: An iPhone switched into Apple's developer mode and tied to an Xcode and Apple ID account, which is required before automation can connect.

Resources

Things they pointed at.

00:20linkKev's iOS Agents (open-source repo) ↗

01:00linkKev's No Code Academy ↗

03:25toolXcode

03:50productApple Developer Program

05:30toolCursor

05:58toolWebDriverAgent

06:26toolAppium

19:10productApify

Quotables

Lines you could clip.

00:36

“We can now automate any task on our phone with 100% accuracy.”

bold claim that opens the video→ TikTok hook↗ Tweet quote

24:16

“That margin is where you can then play around with your price points.”

states the agency business model in one line→ newsletter pull-quote↗ Tweet quote

22:14

“It's tough for these platforms to flag these posts as agentic uploads because the iPhone is being used almost like a real human.”

controversial claim about detection→ IG reel cold open↗ Tweet quote

The Script

Word for word.

Read-along

Don't just watch it. Burn it in.

See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.

The evolution of AI automation has moved so fast through the past few years and it's been wild showing you guys that journey from building end -to -end automations to cloud code skills and now having fully agentic looping. agents and today is kind of like the accumulation of all of my work because what i'm showing you guys today is a open sourced project i'm calling kev's phone agents and essentially what we can now do is get claude to fully control and automate up to a hundred iphones at the same time what people aren't realizing is that since the newer models have gotten so good at computer use we can now automate any task on our phone with 100 % accuracy.

This is seeing cost savings and tokens. We're seeing new capabilities and new workflows that are possible. And this is converting into providing massive value to our clients.

in this video you're going to learn absolutely everything about the iphone farm this is the absolute master class for this and by the end of this video you're going to not just learn how everything works you're going to have your own working iphone farm and you're going to know exactly how to go out there and monetize it and all you need to get started is literally just an iphone and a mac computer really quick before we get started i want to talk about kev's no code academy because we're making a really interesting pivot where instead of teaching you guys ai automations and ai workflows i'm actually going hands -on and helping you guys scale whatever your business is and so if you want hands -on help growing your business join the no code academy and we'll get you going really quick let's get back into the video so today's video is going to go through six core chapters we're going to talk about the exact tool stack that you need to build everything out and scale it to your clients as you move on with this business model we're then going to talk about how it actually works and the different tips that i

have learned on the way building out the project and building out the code base you're going to be using today. And then we're going to go ahead and use the iOS code base to actually set up the phone farm step by step. So I'm gonna show you guys exactly how to do that.

And we'll get our phone connected to that project. live so you guys can see how all of that works and once we have the farm set up we could start talking about the workflows and the automations that you can actually build and how to build workflows properly and make them really reliable starting off with the tool stack now obviously this is a iphone farm that runs on the mac operating system so you're going to start off with the computer and in this use case, I would recommend using a Mac mini or a Mac studio over a MacBook because you want this system to potentially be always on.

And with the Mac mini, since it's not a monitor based computer, you can actually keep this thing on without, you know, hurting the machine. And if you had a MacBook, for example, and you had it plugged in, it might actually drain the battery and ruin the battery. So that's the first point.

The second tool that you need is the iPhone. And with the iPhone, you're going to want an iPhone that's iPhone 11 and up. And that's because the iPhone 11 and lower models can't use the newer iOS operating system and it can't use the newest Xcode updates.

Now, these are very important because that's what actually powers everything. And with the iPhone, what I also recommend is selecting one model size, which will make sense as we get on a little bit later in the video. And so for my iPhone farm that you're seeing here, I'm running iPhone.

12s with the base model as the core phone. Next thing we need is Xcode. Now Xcode is completely free.

It's Apple's developer programming system for you to write React native code. And that's what the iOS code base is written in. And so it makes life really easy because what we're going to be doing with Xcode is we're going to be connecting our iPhone to Xcode.

And once our iPhone is connected to it, then we can do all of this automation. And that leads us to the fourth tool, which isn't needed in the tutorial today, but it's needed for you to scale. And that is the Apple Developer Program.

What the Apple Developer Program allows you to do is turn multiple iPhones into developer iPhones. So for you to actually use any of these automations, you need to go into your phone setting and switch the setting from a normal iPhone to a developer mode iPhone. And then you want to connect your Xcode account to it and your Apple ID account that you've been paying for.

Once you have all of that set up and all those tools ready to go, you can essentially stack any amount of iPhones that you possibly want into the farm. And that's where you want to use something like a connector or USB -C connector. So you can plug in like.

10 different ports to it. So you can plug all your iPhones to it in one shot. Now some other tools that you're going to want are a database and the database is really useful for actually storing data when the iPhone farm goes and does things.

And so if we're doing like lead scraping, for example, and we want to get usernames on Instagram, well, the agent can do that on your iPhone. But it's going to need to then store that somewhere in the database. And so that is why we use the database in this example.

And for you guys to follow along. And when we rebuild the code base, you're going to want to get something like cursor, GitHub. and use an AI model to do the coding.

And so I'm using Claude cursor and GitHub as the stack there. And if you want to follow along with all of those, they will be at the link down below. And that is the tool stack that is actually needed to run things as you play around with it, it'll make a lot more sense.

So now that we have the tool stack out of the way, let's get into the weeds of how this thing actually works. I'm going to do my best to explain this with diagrams. So we're going to have our agent editing really nice.

flow diagrams to show you guys how this works and i'm going to explain the tech stack that is inside of mac's operating system that makes this possible so the first thing that we want to bang out right how does this actually work so we want to understand that apple does not let random software actually control an iphone like there's no tap here on any iphone app that we're automating but there is a loophole and this loophole is through this apple developer testing testing route that this entire farm has been built on.

And that's why you want to set the iPhone that you're working in as a developer iPhone, because then it gives you access to this thing called the WDA service. The WDA service is a web driver agent, which can literally control the iPhone from the web. And so when we connect our iPhone to the web, that is what our cloud agent will be seeing.

And it'll be relaying the actual message. all the way to the phone which is connected on the mac mini and that part is possible through the second portion of the web driver agent called appium and so when you use appium alongside the web driver agent you all of a sudden can control and click around on the screen and even that wasn't good enough for me to make this video it's the fact that claude can also take screenshots of the screen and understand exactly what's on the ui so that it can instruct the perfect button click or button type or whatever the use case is.

And so to paint an example of how this works with a diagram, let's talk about a Instagram DM automation that uses the phone farm. And so in the browser, we can see the phone is ready to go. And in the database, we'll have all of the leads stored.

And so what this automation needs to do is grab the leads from the database, pass it to the cloud agents, have the cloud agents write up a personalized message. for that specific lead and then send that message to the iPhone. Once the message gets sent to the iPhone, the iPhone will wake up, find that Instagram username and then send them a DM.

Once the DM sent, the phone will actually take a screenshot and pass that back to Claude and Claude will verify that that DM has been sent and it'll mark it off as sent in the database and then it'll move on to the next. user in the list. And then once you have that workflow working well, you can then set it up on automation.

And then all of a sudden you have a working automation on your iPhone that is doing something that is costing you no credits as well. So that is essentially how it works all right so chapter four is the fun part we're actually going to build out the phone farm and in this example i'm going to be using my iphone 17 because my real phone farm runs on a different mac mini so let's get into it so in the description down below you're going to find this code base called kev's ios agents and you just want to copy that link and then i'm going to head over to cursor and we're going to open up a brand new window And we're going to clone the repository.

So when we click clone, it's going to paste in that URL on my computer. I already have Xcode installed. I already have all the dependencies installed.

But what you want to do with your first prompt with this agent is you want to ask the agent to install all the dependencies and ask it what it needs to have everything set up. Right. So it's going to probably ask you to sign into Xcode and all of that fun stuff.

You just need to bite the bullet, get all that tedious work out the way. Once you do it one time, you're going to be good to go. Once you get all the dependencies installed, the next message that you need to send in your agent is literally just to launch the UI.

So that's what I'm going to say here. I'm going to say like, please launch the iPhone farm user interface. On the left -hand side, we have the code.

On the right -hand side, we now have the user interface. And this is where we're actually going to register our devices. So at this time, I'm actually going to plug in my iPhone to the Mac Mini.

So the iPhone is plugged in and we're going to now set that down. And on the UI, there's a tab that says register device. And we're going to click on that.

And if your phone is plugged in, it should automatically find your device. And so here we can see that Kevin's iPhone is ready to be set up. And so I'm going to click setup device and here it's going to detect what type of iPhone you have and it'll lay out the actual like screen dimensions and where it'll actually be able to click around on the screen.

And from here is actually where you want to go ahead and do all of your connection tasks. And so it'll automatically check if your iPhone is connected to your Xcode developer account. If it's not connected, then this is going to pop up as a red sign.

uh as well as the wda and the appium actually all needs to be integrated and so once your iphone is connected we're going to scroll down to the bottom here and we're going to authorize xcode to actually go ahead and do all this stuff and we're going to click register and prepare wda and that'll start attempting the web driver agent connection to the iphone and at this time you want to check the iphone because it might ask for some developer permissions and so If it asks for the developer permissions, it's it's a XC test message is what you're looking for.

If you don't get that message, you'll see the error sign that we're getting right now on the screen. And that's OK. Every now and then that'll happen.

You want to just unplug, replug it back in and test it one more time. So after plugging it back in, we can now see the developer status has been checked off. And on our phone now we have the XC test request.

And essentially what I have to do now is just paste in. my code that logs me into this phone. And then it's going to start saying automation running on the phone.

And that's how we can tell that the first part of the connections actually built because our app can now start seeing our screen. And so once that's set up, we can scroll back to the bottom of the page and hit verify with Appium video touch points and accounts, which we'll talk about. right after it gets set up.

So we're just going to give that a couple of steps. And it has now checked off the fact that we are logged into Instagram and TikTok. And on the phone right now, it has popped open Instagram and TikTok to check that it is able to sign into these accounts.

And at this time, the rest of the onboarding for the phone has been checked off. And so we can now tell that this first phone is integrated into the farm. And so this setup is actually never needed ever again with this specific phone in this farm.

All you have to do is keep this puppy connected to the Mac Mini for the time being to move forward and you're going to be money. If you do ever unplug it, though, at one point, the only thing that's going to happen is the XC test request for your password is going to pop up back on the screen. And so all you have to do is type in your passcode and then it'll have all of this integration already built into the phone.

And so that is really the only thing you would have to be paying attention to when you're running a bunch of phones at the same time. So after you hit finish registration, you can now see the live iPhone on this web app and you can start programming the different automations. So I think the best thing to do now is pivot into a chapter where I literally just show you everything on the screen.

And then we're going to go ahead and set up our first automation and how that workflow actually works and how to build new workflows from scratch. and how to integrate new apps that are on your phone from scratch. So let's start off with the devices page.

Now, every single device will have its own unique device page where you can integrate workflows only for that iPhone, or you can use this iPhone to build out workflows for the entire farm. And on the right -hand side, we have the integrated workflows that are already in the code base. So for TikTok, we have a warmup, engagement, create a post, and post pipeline.

These help warm up the account. They help engage with other profiles. They help obviously post to all platforms and post at scale with multiple accounts.

And we did the same thing for Instagram and LinkedIn. And the way that we are building these workflows is by getting the agent to read all of the available pages on a specific application. So for Instagram, for example, what we're seeing on the screen right now is the homepage.

The homepage has so many different locations that we can click around. And what we're converting these different actions on a page is to a thing called touch points. And so if we open up the touch points in the top right hand corner and we navigate to Instagram, we can see all of the available touch points that our agent can use to actually click on the screen.

So when it's building out automations, it's taking screenshots and then it's looking at these different touch points to decide what the next step is. And so if there ever is an automation that runs into an error, this is now how we're able to get the agent to debug any error it runs into in real time without us being there.

And that is the massive breakthrough that we are now seeing with the iPhone 5. With the static clicking workflow before the newer AI models, If an error were to occur, then the automation breaks.

And what these social platforms are understanding right now is that they can change how the UI looks dynamically. And so the homepage won't always look like this. It may look a little bit different and that will mess up these static workflows.

But when you have the agent as the eyes, then this actually becomes a 100 % working workflow no matter what. And that's what creates such an advantage for us as the agency. Aside from the devices page, we also have the automations tab.

And in the automations tab, you're going to be able to see all of the live automations that you have set up with your agent. And on your Mac mini and on Apple's operating system, there's something called the launched agent service, which is already built into Mac OS, which is literally how you can set up automations locally without paying for any cloud provider.

with the Mac mini, since it's staying on, it can call this automation and then use the iPhone to complete the task. Now, some other things that we can do on this page is actually control the iPhone ourselves, and we can just click around or we can use these quick action buttons. And so I don't necessarily use this ever unless I'm working with the agents to build out a workflow.

And so now guys, before we start building our first automations, I want to talk about the two different types of workflows that the iPhone farm unlocks for us. The first one is much simpler. So it's a static workflow that you already know is going to be working.

And the goal is to turn everything into a static workflow, even though that's not necessarily possible in all of these different use cases. that should still be your North Star. And the reason why is because you're not using any AI credits when you're running these automations at scale.

Because once the agent understands where on the screen it needs to touch to actually complete a said workflow, that is then considered complete. The only time that would then have to change is if the UI changes, and that's what I was explaining earlier. But where the real massive unlock is, is using a really fast, intelligent, and cheap model like Haiku 5 .5 as the eyes and the brains behind a said workflow.

workflow. And this allows you to do really dynamic, multi -step, agentic workflows on an iPhone. And there's certain things that that unlocks that weren't possible through APIs or through web provider applications, like sending iMessages at scale, like interacting on TikTok and Instagram, like sending DMs on Instagram, like all of these different sub use cases.

And now you're able to do this at a low cost, at a high speed. at scale and that is really the massive unlock is that you have this dynamic agent that is able to actually take a screenshot and through this looping workflow it's going to know whether or not the task is completed and so let's start things off by showing you guys a static workflow and then we're going to build a dynamic workflow together and then we're going to finish today's tutorial off by setting that dynamic workflow on an automation that runs without us so on the screen we got some of the automations for tick tock that are ready to go and in this example let's say we want to run the 4u warm -up tick -tock automation which is the static workflow it's going to pop open the for you page it's going to scroll it's going to to like comment and save on some of these posts.

So I'm just going to tell the agents to run that so you guys can see now one side note is that this code base has been wired and programmed for my iPhone's dimensions and so I recommend even though we have the workflows built in I recommend actually testing them out to see if they work on the iPhone that you have connected to the farm because the dimensions might actually mess things up and then your agent will have to rewire all of the different things but right now my hands are completely off the screen not touching anything and we see that the agent has successfully done its first scroll and occasionally it'll end up liking commenting and saving it doesn't necessarily need to do that on every single scroll.

And on this specific video, we saw it give it a like. So we're going to let this kind of go for a bit. All right, awesome.

So I gave it about a minute there to scroll just to see whether or not it's working. And in this case, we can see that it's good to go, right? So we can quickly pivot to a dynamic workflow.

Now, in this dynamic workflow example, let's talk about lead scraping. And so what I'm going to get the agents to do is go on my Instagram page, find all of my followers by heading to my follower. by heading to my follower list and then scraping all of the usernames and storing them in a database.

For me to do that exact same follower scrape on Appify, that would have costed me around $2 per 1 ,000 results. And with 8 ,000 followers, that would have costed me like $16. And in this example, it's going to cost me absolutely nothing except for Haiku, which is going to use the phone, which will literally cost me pennies.

And so that's the prompt that I'm going to send in here as a dynamic workflow that the agent and I are going to work on. So my prompt here is that we're going to stop the current workflow and build a dynamic one using Haiku 5 .5 that goes to my Instagram, then heads to my follower list, goes and scrolls on the follower list, takes notes of all of the followers, and then stores them in a database.

So here's a multi -step agentic workflow, and I'm going to tell the agent to learn how to do this on the fly. You're going to see exactly what I'm going to see. I'm going to send this prompt and we'll see what happens.

After that fast forward, I kind of just sat here and we can see the agents going through the scroll. It's using Haiku 5 .5 to do this. And it's capturing all of the usernames, right?

So I'm going to now let this run for a minute, minute and a half, right? We can see how we can already improve the workflow. It's scrolling a little bit too aggressively.

It's maybe missing some of the usernames and little things like that. But now let's, you know, imagine like this workflow is working and we're grabbing, you know, 100 usernames at a time. And this workflow can now...

instantly be turned into an automation that the agent launches at set schedules and so that will be the next task that we would tell the agent we're going to be like okay this thing's able to get a hundred different uh usernames every single time it runs and we only need about you know 100 usernames a day to send dms to so this thing can run twice a day to get more than enough leads that we need right so that's going to tell the agent here and like run this twice a day And the next time that we'll run it, let's say it will be like 7 .27 PM so that you guys can actually see the automation launching completely by itself as well.

So that's what I'm going to now tell the agent. All right. So the prompt here is basically saying I want the next automation of this 100 lead capture to launch at 7 .28 PM Eastern time.

And I want that to happen because I want to see. the automation launch for the first time by itself so that i know that the entire pipeline is actually working so i'm going to send that message in fingers crossed that that works well And that will be that.

And then we can move on to the actual like business use cases and like stuff that the stuff that businesses are actually looking for that you can sell with this automation farm. Okay. Message from the agent says the next run is in at 728 PM.

So we can head over to the automations tab here and actually view whether or not we have things set up. And so, yeah, if we head over to the queue on the schedules, we have one at 728 AM and one at 728 PM. ready to go completely active.

That looks good. It is now 728 PM. So hopefully if we head back to the device, we can see that it's actually there.

So let's go open device and it has already navigated back to the following page. So we can now verify that this automation is working. So while that's running, you guys can kind of watch while that's happening.

I want to talk about like the different business use cases and just like, there's such a rabbit hole here because the amount of use cases and the opportunities are kind of growing. Right. And so like an example of this that I want to start off with is just the Apple based workflows that you can build that don't exist anywhere else.

So that is like iMessage marketing, for example. If you wanted to get an agent to do iMessage marketing without an iPhone farm, there's services that charge like $200 a month for this. But now you can literally have your agent and give it its own Apple ID, give it its own iMessage, its own phone number, and it can just use the iMessage in there.

And so that can be used for like e -commerce. brand notifications. It can be used for like coaching.

It can be used for software. It can be used for updates, events, like the list goes on. And that's just like one minor example.

Another one is posting at scale to different like fan pages or brand accounts. And this can apply for musicians. And so like the musicians that we work with, for example, they're not necessarily looking for videos to go.

viral or make sales. They're just really looking for impressions on their songs. And so you can put like a trending video and then put their song behind the trending video and then have, you know, 200 different TikTok accounts, 200 different Instagram accounts posting.

to these different fan pages and you can warm up these fan pages in the fans niche by doing the engagement workflow as well so you can have warmed up fan pages in your fans niches and post on their behalf and so if you're generating them like 10 20 million views of their song they get a little bit of pay for their song they get a little bit of exposure they get some potential ticket sales to their concert like that's an entire rabbit hole and that same type of posting pipeline can be used for winning e -commerce products So if there's a winning e -commerce brand or a viral product, you can now grab that product, work with that company or build it yourself and then repost the winning product video to all of these external accounts at the same time again.

And since they're warmed up, they're going to be able to get that reach. And it's tough for these platforms to flag these posts as agentic uploads or as AI based uploads because the iPhone is being used almost like a real human. versus having a computer use workflow that can do the same thing.

It's much easier for these platforms to identify those as AI -based workflows. So that's really cool as well. And like the one example that you're seeing right here is a lead scraping example.

And so what I want to talk about in this example is not really lead scraping. It's not really lead sourcing. It's cost reduction.

So if I see a workflow that exists in my current tech stack or my current client's tech stack, I can literally think about. How can I reverse engineer it inside of the iPhone farm to eliminate the cost? So if we're doing like follower lead scraping, that could be like $20 a month.

And then lead enrichment, that could be like another $20 a month. And then like sending connection requests on LinkedIn, that's another $20 a month. And these are all different workflows.

can all be eliminated with the cost. And so that margin is where you can then play around with your price points. So that guys is the iPhone farm masterclass.

If you want any personal help, hands -on help to actually set this up, you can either book a call or join the no code Academy. This now lets me show you guys all of the individual workflows that I've been working on in the farm as well for different businesses. And so that's really exciting.

So hope you guys enjoy. I'll see you guys in the next one.

The Hook

The bait, then the rug-pull.

Kev Builds Apps claims Claude can now fully control up to a hundred iPhones at once, and spends the next 25 minutes showing the Mac mini, the open-source repo and the workflows behind it.

Frameworks

Named ideas worth stealing.

16:00model

Static vs dynamic workflows

Static: replay known touch points, zero AI cost, breaks if UI changes
Dynamic: model loops on screenshots, adapts to UI changes, costs pennies on a fast model

Aim to convert every workflow to static once it works, and use a cheap dynamic agent for steps that need judgment or self-repair.

Steal forany automation with a mix of repeatable and judgment steps

05:45list