How I built GuestLens around planning a wedding
This is the build side of GuestLens. The reasons behind it, and how it works for guests and hosts, are in We'll miss half of our own wedding, so I built an app for our guests' photos.
I'm an iOS developer, not an Android developer, and until this project I had never done proper web development. That mattered, because GuestLens needs a backend, and I didn't want to run and host an app on top of planning a wedding. It's why I spent close to a month looking for something I could just use instead. When nothing worked, I built it.
Planning and building
I sat down with Claude to sanity-check whether a realistic timeline was possible as a solo iOS developer. Then I worked out the architecture before writing any code: a hexagonal architecture like the one reHydrate uses, with a focus on testability. When an app has to be done in a couple of months, you need tests to make sure you haven't broken anything you already fixed.
From there I went spec-driven. I dedicated real time to writing a detailed specification for each feature, then handed the implementation to AI. Then AI handed it back to me to QA and test on real devices. AI helped write the code, but the architecture and the feel of the app are mine.
I also had to keep the feature set small, because the launch deadline was my wedding and that can't move. If the app wasn't ready for the big day, I'd have had to use one of the annoying web views I set out to replace.
The first working version took about a month. I spent another month polishing it until I'd be comfortable handing it to a paying customer, instead of stopping at a rough first version. The stack is Swift and SwiftUI on iOS, Kotlin on Android, and Firebase.
I tested it with my parents and friends, and hired QA testers to get feedback from people with a range of tech skills.
Android was the hard part
I'm not an Android developer, although I have written code in Kotlin and Java and created one Android app for my uni. Creating an app from scratch without knowing the lifecycle was challenging. However, a guest list often has both iOS and Android users, and our wedding is no exception, so GuestLens had to support both. Because of the deadline, I cut features from the Android app and made it guests-only. Hosting events on Android should follow shortly after Christmas.
That being said, the hardest part wasn't the code or understanding Android's paradigms. AI was a great help with that. It was Google Play, which requires a minimum of 12 testers using your app every day for two weeks before you can start a public testing group. From talking to some Android developers and reading forums, getting testers to use an app every day for at least two weeks is hard enough for any app.
Having to use the app every day for multiple weeks goes against the main use case of GuestLens. I want guests to use it on the day and the next couple of days, not every day. However, it turns out this has become a business: you can hire testers relatively cheaply to use your app every single day, for as long as you need.
Going through this made me really appreciate the iOS experience. The iOS community hates how long it takes Apple to review apps. But there isn't a hard requirement to have a set of users using your app every single day for a certain number of days or weeks before you can get it on the App Store. Apple does this on their side, and they pay people to test your apps. Apple does it for every release, though, while Android only requires its testing once. So each approach has its positives and negatives.
Where it is now
GuestLens is live today on the App Store and the Play Store. It isn't a beta or a waitlist. Right now my focus is getting customers, to see if this app has legs to stand on. I'm also constantly getting new ideas for features the app should have, but there will now be a change freeze until the wedding so I don't break something I need on the day.