September was supposed to be quieter
An App Store rejection, an approval, a shop launch, a builder café, 3,000 people on X, two new domains and the beginning of an AI sidekick.
September was supposed to be the month I settled back into something resembling a normal routine.
Continue reading
I went back to my 9–5 on 1 September, which in my head meant the building would naturally slow down a little. I would choose one or two projects, work on them during sensible hours and perhaps even become the sort of person who finishes one idea before beginning another.
By 3 September, Lumarii had been rejected by Apple. By 7 September, it had been approved and was ready for distribution.
That set the tone for the rest of the month really: a lot of emotional whiplash, far too much coffee and every tidy little plan immediately growing another branch.
Back to the 9–5 and straight into an App Store rejection
Lumarii is my quieter take on journaling and reflective to-do lists. It had already taken a lot of thought to get it to the point where I felt comfortable sending it to Apple, so opening September with a rejection was not ideal.
App Store rejections always feel strangely personal, even when you know they are not. You have spent weeks or months staring at every tiny part of something, then somebody else looks at it for five minutes and sends it back.
A few days later, Lumarii was approved.
Nothing about the app had suddenly become more meaningful because it had passed review, but seeing it ready for distribution still felt like a proper win. Cozy Pom, MochiMoney and Lumarii were now all real products rather than folders full of Swift files living on my Mac.
Of course, getting an app approved does not come with an audience. Apple does not open a secret tap somewhere and start pouring users into your analytics. You still have to explain what you have built, find the people who might care about it and somehow convince them to stop scrolling long enough to take a look.
September became the month I started thinking much more seriously about that side of building.
Three apps, built around a full-time job
On 16 September, TheCozyDev reached 3,000 followers on X.
I still find that number slightly ridiculous. The account began as somewhere to share the things I was making after work. There was no strategy behind it, no big launch plan and definitely no expectation that thousands of people would eventually follow along.
Reaching 3,000 gave me the push to finally write more honestly about how all of this started. The result was Three apps, built around a full-time job, an article about building Cozy Pom, MochiMoney and Lumarii during evenings, weekends and whatever small pockets of energy were left after a day in credit control.
I wanted it to feel like the real version of the story. Not the polished version where every idea arrived fully formed and each app sailed through review, but the version involving rejection emails, overcomplicated settings screens, external drives full of Xcode projects and a worrying amount of coffee.
The article also made me stop and look at how much had changed. TheCozyDev had started with iOS apps, but it was already becoming something much bigger than that.
The Cozy Shop finally became real
A huge part of September was spent building Ship Your First App Kit.
I wanted to create the resource I would have needed when I started. Not another enormous course promising to turn somebody into a founder overnight, but something practical and calm for the person sitting there with an app idea and absolutely no idea what order to do things in.
The kit became a £24.99 one-off product covering the journey from an early idea through planning, building, testing, App Store submission and finally getting something into the hands of real people.
Creating the guide itself was only one part of it. I also had to build the shop, write the sales page, design the cover, create the product graphics, plan the launch content and work out how the private web companion would operate.
On 12 September, I settled on Stripe Managed Payments. After spending most of my development time inside SwiftUI, I suddenly found myself learning much more about authentication, payment flows and how to make sure only paid customers could access the right content.
I used a day off from the 9–5 on 18 September to keep working through the backend. A day off probably should not involve authentication logic and product copy, but apparently that is what counts as relaxation now.
The shop opened on 21 September.
ShipKit was finally real, purchasable and sitting on thecozydev.co.uk rather than existing as another increasingly complicated collection of notes. I even started looking at whether a physical paperback version could work through Amazon KDP, because launching one version of something was apparently not enough.
The website changed with it too. TheCozyDev was no longer just a homepage pointing towards my apps. It was beginning to become a proper home for the things I build, the things I learn and the resources I want to share.
Learning a completely different side of building
Softr became a bit of a testing ground during September.
Alongside the private companion area for ShipKit, I started putting together an App Store Connect Companion. The idea was to make the process of tracking apps, releases and all the other slightly scattered parts of App Store work feel less overwhelming.
The database even got the extremely imaginative name TheCozyDev App Store Companion.
Working on web products has stretched a completely different part of my brain. In an iOS app, I am used to thinking about views, state, navigation and what happens when somebody taps a button. Building paid web products added accounts, permissions, secure access, databases, checkout flows and several new opportunities to accidentally lock myself out of my own work.
None of it was especially glamorous, but it opened up a new side of TheCozyDev. I was no longer only thinking about what could live inside an app. I was beginning to look at products, tools and little connected experiences across the web too.
My MacBook was also rapidly running out of storage, with more and more projects being pushed onto an external drive, so I ordered a Mac mini to become the main build machine. It was one of the more sensible purchases made during a month that also involved buying multiple domains.
Pocket Town started becoming a real world
Pocket Town also moved forward properly in September.
The idea began as a way to help children learn about money and maths without turning either subject into a lecture. During the month, I locked down the first version of the product: the age groups, the town, the mascot, the ten core games, the XP system, parent and child accounts, subscriptions, retention and the safety side of the experience.
I worked through the main onboarding flows for parents and children, the Pocket Town home screen, Coin Café, reward moments, room customisation and Home Missions that a parent could approve.
Locking an MVP sounds like you have made something smaller. Pocket Town somehow managed to feel bigger once I had decided what belonged in it.
Still, it was the useful kind of bigger. The project stopped being a vague idea about financial literacy and started becoming a world with actual rules, places and reasons to return.
The side projects continued multiplying
A normal person might have looked at three apps, a shop, a web companion and a children’s game and decided the month was full.
I did not do that.
I started sketching out a freelancer invoicing product based on the part of my brain that has spent years working in credit control. The early description was basically a credit controller in your pocket: something to help freelancers send invoices, keep track of what is owed and follow up without feeling awkward about it.
The gaming journal idea also became Questory, a calmer place to track quests, collections and the games you want to spend more time in. I wanted it to feel less like a database and more like a town you would actually want to hang around in.
Later in the month, Medfolio appeared too. The idea was a calmer place to hold medical records, history and the information that becomes almost impossible to remember when you are actually sitting in front of a doctor. I began exploring the name, logo, icon and early mock-ups.
I also spent time thinking about a concept I called designing in silence. Most tools react to an idea immediately, but I kept wondering what would happen if a piece of technology deliberately did nothing for a while. No instant critique, no list of features and no AI turning a half-formed thought into a product roadmap before it had been allowed to become strange.
The idea was to give thoughts somewhere private to sit before asking technology to get involved. I still do not know exactly what that becomes, but September reminded me that not every idea needs to turn into an MVP by the next morning.
Then I accidentally started building a café
On the same day ShipKit launched, I registered cozycafe.dev.
The original idea was simple: a builder directory where everyone pays the same £1, chooses a drink and gets a little menu card inside the café.
No boosted listings. No paid hierarchy. No leaderboard to climb. Everybody gets a seat at the same table.
Naturally, the idea did not remain simple for very long.
The directory started becoming a warm, slightly strange retro café with paper menus, drink icons, shiny plastic-covered cards and room for builders to discover one another without feeling like they were being ranked.
I moved from early Framer experiments into Figma and started building it around an existing Supabase project. Menu cards needed names, bios, product links and optional project details. Drinks needed illustrations. The café needed tables, menus, doors and enough tiny details to make the whole thing feel like a place rather than another directory grid.
By the end of September, I had created assets including the black coffee and lavender latte, worked through the visual direction and started thinking about the stories and secrets that could live inside the café.
The doors were still closed, but the room behind them was beginning to exist.
The numbers went up, then the reach went sideways
TheCozyDev was at around 2,800 followers by the middle of September and reached 3,000 on the 16th. I also set myself the slightly unhinged goal of eventually reaching 10,000.
Growth looked good from the outside, but the actual reach became much less predictable.
By 28 September, verified impressions were sitting at around 1,800 a day, with roughly 20,000 across seven days. Those numbers were not disastrous, but they were frustrating when compared with the size of the account and the amount I was posting.
It was a useful reminder that followers and reach are not the same thing. A number on a profile can go up while individual posts still feel like they have been placed quietly in a cupboard.
I kept experimenting anyway. I shared more of the real building process, joined the 5-4-3-2-1 indie developer challenge on TikTok and X and admitted that my current obsession was simply building.
Some posts travelled further than others, but the best part of the month continued to be the people around them. The conversations became more interesting, the ideas got stranger and TheCozyDev started feeling less like an account broadcasting updates and more like a little corner of the internet with familiar faces.
A growing corner of builders
September included quite a few small community moments that meant more than the numbers.
Guy launched PlatPost, and I somehow became its first subscriber. I spent time testing it, giving feedback and watching a friend put something new into the world.
Adeel was closing in on 2,000 followers, so people gathered around to give him a push. I claimed spot #004/100 on Vibe-code.ing and joined theNetwork.builders, adding a few more connections between my little collection of projects and the wider builder community.
None of these things were enormous announcements on their own. Together, they made the month feel much less solitary.
Building around a full-time job can be lonely. Most of the work happens when the rest of the house is quiet, and a lot of the progress is invisible until something finally ships. Having other people around who understand why a working login screen can completely transform your evening makes a real difference.
Cozie arrived on the final day
On 30 September, I created Cozie.
Cozie was not meant to become another coding agent. I already had plenty of ways to generate code and plenty of opportunities to create new bugs all by myself.
I wanted her to become the operational brain behind TheCozyDev: something that could keep an eye on projects, surface useful conversations, help with content research and stop every unfinished idea from disappearing into a different notes app.
The first routines included a 9am Morning Brief covering builders, iOS, AI, product design and more experimental bits of technology. I also created a weekday Loose Ends and Tomorrow check-in to separate what needed attention now, what was coming soon and what could safely be parked.
It was still an experiment, but it felt like the beginning of a different relationship with AI. Less “build this entire thing for me” and more “help me see what I am already juggling.”
On the same day, I bought cozylab.si.
There was not yet a sensible plan for the domain. I just knew I wanted somewhere for the smaller tools, experiments, components and ideas that did not belong inside one of the main apps.
Buying a domain before fully understanding what it is for has become something of a TheCozyDev tradition.
The part that was harder to write about
I do not want to dress September up as one long productivity success story.
The month began with me returning to my 9–5 and ended with me needing time off again because of post-hospital pain. My neurosurgery pathway was reopened, with another letter expected in the following six to eight weeks.
There is a strange space between being incredibly proud of what you have managed to do and realising that your body does not care how many things you shipped.
A lot happened during September, but none of it happened in perfect conditions. It happened around full-time work, health problems, appointments, tired evenings and days when opening Xcode felt much harder than posting the finished screen afterwards made it look.
I am proud of the output, but I am also trying to get better at recognising the cost of it.
Rest cannot only be the gap where I plan the next product. A day away from work cannot always become an extra development sprint. I am still learning where the line is, and I am still not especially good at stopping before my body makes the decision for me.
What September actually became
September began with an App Store rejection and ended with three released apps, a live shop, a nearly finished café, a growing community, the beginnings of Cozie and more active projects than any sensible person would recommend.
Ship Your First App Kit went from notes to a real product. Pocket Town gained a proper shape. The App Store Connect Companion moved into Softr. Questory got its name. Medfolio started to appear. CozyCafe went from a domain to a place with tables, drinks and closed doors waiting to be opened.
TheCozyDev reached 3,000 followers, while also teaching me that an audience cannot be measured by one number. I wrote more honestly about building around a full-time job, joined more of the builder community and kept sharing the parts that usually happen before a polished launch post appears.
Not everything worked. Some things were rejected, some remained half-built, reach dropped, my body forced me to stop and several ideas arrived before I had anywhere sensible to put them.
But somewhere between the apps, the shop, the café, the lab and the slightly overworked AI sidekick, TheCozyDev began to feel like more than a place where I post development updates.
It started to feel like its own little world.
I am not going to pretend October will be quieter. There is a café door waiting to open, Cozie has only just switched on, ShipKit still needs more people to discover it and there are far too many ideas living in my notes.
At least the coffee is ready.
