Turns out there was more to talk about in this journey. Not that surprising I guess 😅. Today I’ll go through one aspect that isn’t talked about as much, and where these tools clearly still struggle. At least based on this specific experience.

The pre-launch reality 🛠️

Producing code is one thing, and these agentic tools, especially Claude Code, are awesome. Guiding you through setting things up in a portal is a whole different game.

And before going live (more on that below), you have to set up quite a few things outside the code. Different Firebase integrations, RevenueCat for subscription management, AdMob for ads support, and most importantly Google Play Console to be able to distribute the Android app. Also the Apple Developer account, but we’re not there yet.

Where the tools struggle 🤔

I don’t know if it’s because these portals change so much and so fast, and there’s so much conflicting information available online, but this isn’t the type of help where these tools shine.

Yes, they can point you in the right direction, but don’t expect the same level of detail and accuracy as with producing code. Menus change names. Rules evolve. Processes change. As opposed to code where I had almost no back and forth, here there was a lot of validation whenever I got blocked.

Screenshots help a lot 🖼️

Screenshots can be very helpful, because, even though it’s clearly new information for the model, it’s able to interpret what it sees using the context it already has and guide you in the right direction. It does a great job recognising the UI and putting it together with the broader context of what we’re trying to achieve.

This was especially important during the setup for subscriptions in Google Play as I had never done that before and was not super confident in those flows and what I was really enabling.

A real example: subscriptions setup 🧪

Getting into those subscription flows was actually pretty frustrating. The process to integrate with RevenueCat was not straightforward and slowed me down quite a bit.

To not waste my Claude Code tokens on that type of portal setup work, I initially used ChatGPT to help me. It suggested setting up subscriptions in Google Play first, but because I hadn’t integrated the SDK yet, the app didn’t have billing permissions in the manifest. That meant the subscription menu was always locked.

Then it told me I had to promote the app from internal testing to closed alpha because Google had probably changed the rules. This forced me to create a store listing, fill all the content declarations, and set up things I wasn’t planning to do yet, which slowed me down a lot. Turns out that was not the reason for the UI to be locked.

In the end, the fix was simply to follow the RevenueCat documentation and set up the SDK first. The moment a build with the SDK hit the store, the whole flow unlocked immediately.

It was frustrating at the time, mostly because it broke the fast pace I had during the coding phase. But it also forced me to get a lot of the store setup done much earlier than I probably would have otherwise. So I’ll take the positives from it.

The pace trap ⏱️

Regardless of whether you’re using these tools for that part of the setup or not, the obvious takeaway here is the pace aspect of the project. During the coding phase you’re going so fast, with these tools producing code and features so quickly. It’s very easy to get too excited and believe you’re going to have a final product in a very short amount of time.

When these more manual setups come along in the process, you will slow down considerably and it will even feel boring at times. Be mindful of that when you set a very ambitious release target. The pace is not constant until the end.

Getting to 1.0.0 🎉

You’ll get there though, as I did. Version 1.0.0 of my app is ready to release on Android 🎉.

For iOS I will need to go through the same process of setting up my developer account, pushing the app, filling all the mandatory content declarations, and setting up the subscription packages I will have in the app. But I wanted to test Android first as I’m more familiar with the process and already had a developer account set up as well.

Want to help test? 🙌

For an indie developer account on Android though, Google requires the app to be live for at least two weeks in internal testing with at least 12 users.

Going through that process now but still missing a few users. Want to join the internal testers group and help me validate that everything is working as expected?

Send me a message with the email you use in your Google Play Store account and I’ll add you in.

Wrapping up ✨

That’s it for today. If the Apple setup ends up being substantially different, for better or worse, I’ll share a bit more. In the meantime, back into building more functionality in the app.