It all began with a tiny microphone. I had an Insta360 Mic Air plugged into my Samsung Galaxy A55, and one simple wish: press a button and get a clean, uncompressed WAV file. No account, no cloud, no fuss. The Play Store, that vast and noisy bazaar, couldn’t grant my humble wish. So, like every stubborn engineer before me, I decided to build it myself.
I didn’t walk this road alone. My companion was Claude, an AI from Anthropic that lives in my terminal through Claude Code. I’d describe what I wanted, and Claude would write the Kotlin, run the tests, and come back with questions I hadn’t thought to ask. What happens when the phone runs out of storage in the middle of a sentence? What if Android kills the app while the file is still open? What if the user unplugs the mic mid-recording? I was the one with the idea, the phone, the mic and the final word. Claude was the tireless builder who never once complained about my 2 a.m. ideas.
The first commit landed on July 26th. Over the next ten weeks the app grew from a stubborn prototype into something I could trust. It records in the best format the microphone can really deliver, keeps recording with the screen off, survives crashes without losing the audio, and splits long recordings into neat 30, 45 or 60 minute files. More than 800 automated tests stand guard over it. If you want to see what’s under the hood, the full technical breakdown is on the Projects page.
Then I met the gatekeeper. Google, in its wisdom, requires new personal developer accounts to run a closed test with at least 12 testers for 14 days in a row before an app may see the light of production. Twelve testers. I have friends, but I don’t have twelve friends with Android phones who are willing to open a voice recorder every single day for two weeks.
So I did what every desperate developer does: I paid a testing service. For about the price of a pizza, fifteen strangers from around the world joined my test. Their dashboard proudly said “15 testers”. Google’s statistics told a humbler tale: three, four, maybe six phones on a good day, scattered across five countries. I watched those numbers every morning like a farmer watching the sky.
But the test wasn’t wasted. Using the app every day myself, I found what was missing. Version 1.1 added Share, Save a copy and Rename, because a recording you can’t get off the phone isn’t worth much. Version 1.2 added a Send feedback button and a small “What’s new” card. Two updates, shipped in the middle of the test, through Google’s review each time.
Halfway through the test, the testing service sent its report, along with ready-made answers for Google’s production form. I read them twice. They described surveys I had never run, “audio enthusiasts” I had never reached out to, and a dark mode I had supposedly “implemented” (the app had followed the system’s dark theme since day one). It would have been so easy to copy and paste. Instead, Claude and I sat down and wrote the truth: 3 to 6 phones a day, no crashes, two updates, and what I had actually learned. If the gate was going to stay closed, at least it would be closed on an honest man.
I applied on October 30th, at one in the morning. Google said it could take up to seven days. It took less than one. “Congratulations! Your app has been granted Google Play production access.” I read that email more times than I’d like to admit.
On November 2nd, Bukin Mic Recorder went live on Google Play, in 178 countries. The very next day I sent it a fresh coat of paint: a warm and soft redesign called version 1.3.
What did this road teach me? That the code is the easy part. The hardest parts were the waiting, the paperwork, and the temptation to take a shortcut that isn’t yours to take. That a gate is also a teacher: the 14 days forced me to live with my own app, and it’s better for it. And that a partner who never sleeps is still only as good as the questions you bring to it.
The road was long. The destination fits in my pocket.
Get it on Google Play · Read the technical story · Source code on GitHub
