Skip to content

Fixing 1-2-3 Math with Jules: How Google’s AI Coding Agent Revived My Broken App

Published:
3 min read

In this post, I share how I used Jules, Google’s asynchronous coding assistant, to fix a broken educational game project I had abandoned. The app, 1-2-3 Math, had various issues—broken buttons, language-switching bugs, and general logic problems. I decided to test Jules as a dev assistant and was positively surprised.

🤖 What is Jules?

Jules is an experimental tool from Google Labs. Think of it as a GPT-based developer agent that understands your repo, can reason through issues asynchronously, and even make pull requests with real code suggestions. You just describe your problem in natural language, and Jules builds a step-by-step plan, implements the changes, and pushes commits back.

📸 Screenshot of Jules plan and code fix

Jules plan and code fix

💡 Why I Used Jules

The app was failing in several critical ways:

I didn’t want to spend hours debugging the front-end logic, so I gave Jules a try.

⚙️ How Jules Helped

Jules broke the task into a detailed checklist:

Each step was tracked, executed, and documented. In the final commit, Jules:

📸 Screenshot of merged Pull Request on GitHub

Merged Pull Request on GitHub

✅ Benefits of Using Jules

🔚 Conclusion

Jules isn’t just a toy; it’s a powerful productivity multiplier for developers. For projects that are small, experimental, or stuck in debugging limbo, it feels like having a reliable junior dev always available. For me, it helped bring 1-2-3 Math back to life.

I’ll definitely keep experimenting with Jules—especially for prototype-level work or legacy bugs.


📸 Screenshot of Jules homepage

Jules homepage


Edit on GitHub

👋 ¡Hola! Pregúntame lo que quieras sobre el blog
🤖 AI Assistant