I Built an Android App Without Writing a Single Line of Code (Using AI Agents)

I Built an Android App Without Writing a Single Line of Code

AI tools are changing how we build software.

Recently, I experimented with a different development approach: building

an application using AI agents instead of writing the code myself.

The result was a fully working Android application with a Spring Boot

backend.

The project is called Sub Tracker.

What is Sub Tracker?

Sub Tracker is a mobile application that helps users manage and track

their recurring subscriptions and regular payments.

It allows users to:

• track active subscriptions


• receive reminders before payment dates


• manage recurring expenses


• access premium features through in-app purchases

The application is currently in Google Play Closed Testing.

AI-Assisted Development

One of the main goals of this project was to explore AI-driven software

development.

Instead of writing most of the code manually, I used AI agents to

generate and iterate on the implementation.

Tools used in the development process:

• Gemini


• Claude


• Antigravity (agent orchestration)


• Google Stitch for UI design

My role was mainly to:

• define the problem and requirements


• guide the AI agents


• review and iterate on the generated code


• design the overall architecture

This experience showed me something interesting:

AI is not ending software development. It is changing how developers

work.

Developers are gradually becoming system designers and orchestrators

rather than just people who write code.

Tech Stack

Mobile

• Kotlin


• Android SDK


• Jetpack Compose


• Material Design 3

Networking

• Retrofit


• OkHttp


• Kotlin Coroutines


• Flow

Dependency Injection

• Hilt (Dagger-Hilt)

Monetization

• Google Play Billing


• AdMob

Backend

• Java


• Spring Boot


• REST API


• MongoDB

Architecture

The mobile application follows the MVVM architecture and is inspired by

Clean Architecture principles.

Layers:

• UI Layer – Jetpack Compose screens


• ViewModel Layer – UI state and logic


• Data Layer – Repository and API services

The application communicates with a Spring Boot backend via REST APIs.

Monetization

The application includes two monetization strategies:

• AdMob advertisements


• In-app purchases using Google Play Billing

Project Repositories

Android Application

Current Status

The application is currently in Google Play Closed Testing.

The testing phase focuses on:

• application stability


• monetization flows


• backend performance


• user experience

Final Thoughts

This project was an experiment to see how far AI-assisted development

can go.

One thing became very clear to me:

AI is not replacing developers.

Instead, it is transforming the role of developers into something closer

to:

• system architects


• AI orchestrators


• product thinkers

And honestly, this is just the beginning.