my attempt to.... series : ep 2

18 Nov, 2025

I’ve always loved building things for myself — tiny tools, random ideas, and those mini “attempt” projects I keep writing about in my blog. Earlier, I wrote about my attempt at starting a series EP. Today, I want to talk about something much bigger : my attempt to build a full web app for myself. This wasn’t just another project. This was a whole journey about how my ideas evolved, how I figured out what I really wanted in a personal app, and how I actually learned the skills to make it happen.

🤖 When One AI Started Writing for Another

I wanted a web app that wasn’t generic — not another todo app or basic dashboard that everyone builds. I wanted something useful specifically for me, something that matched my daily routine, my habits, my aesthetic, and even my friendships. But here's the twist: I didn’t want to sit and write heavy code line by line. I wanted to build this with AI — but smartly.

At first, I explored generative AI tools to generate the entire codebase, but I realized they weren’t efficient enough to build a full-fledged working app from scratch. So instead of relying on them fully, I learned how to use vibe-coding tools like Cursor and Windsurf (I used Windsurf in my case). These tools are literally built for such workflows. I watched a ton of tutorials, learned how to prompt properly, understood the typical issues that show up in app development, and slowly got the hang of the whole process.

⭐ Features I Dreamed of Having

1. A personal dashboard
2. A daily planner
3. A water tracker
4. A money tracker
5. A space to share favourite music with my frd
6. A shared gallery where we can store rare aesthetics
7. A shared todo list
8. A shared scribble card (just for fun!)

I asked Gemini to generate a complete plan, tech stack, and execution strategy for this entire project. Then I fed all those prompts into Windsurf to start the actual build.

🧪 Lessons I Learned

Things didn’t go smoothly — of course they didn’t. I learned that:

1. You shouldn’t build an entire app in a single AI chat
2. Old context makes the AI confused
3. Too many tokens make responses inefficient
4. Every feature deserves its own new chat for clean implementation

And yes… I hit usage limits. So I switched accounts midway (tech problems require tech solutions 😭). Eventually, I built the entire backend, and then moved to the second account to build the complete frontend.

🎨 Designing the UI Exactly the Way I Wanted

This was my favourite part and also the most time-consuming. I wanted cute doodles, buttons that actually matched the theme, and a layout that looked aesthetic but still usable. My app didn’t need huge storage, so I used Firebase for authentication and storage — perfect for a lightweight personal app.

Then came the painful part: Making the whole thing responsive for mobile. I had to modify the entire frontend so it wouldn’t break on mobile screens. It took hours, but honestly, I loved seeing the UI finally fit perfectly.

🚀 Deployment — The Most Satisfying Moment

And finally — deployment. I used Netlify, and it literally took just a few minutes. Just like that, my scrapbook-style personal web app was live.

I had spent nearly 4 full days (around 9 hours each) — 2 whole days went just into finding the right doodles and setting the UI vibe. But it was totally worth it. Today, I use this app daily — for my work, my routine, my health, and even my pockets hehe.

This was my attempt at building my own Scrapbook Web App. And honestly? It worked out so well.