EchoType

Key Features

Tech Stack

• Java
• Spring Boot
• React
• Vite
• OpenAI API
• Axios
• Postman

Why I built EchoType

EchoType started as a classic "what if I just..." kind of weekend idea. I had a break between courses, was curious about OpenAI's Whisper model, and thought, why not build something that turns audio into text? It wasn’t meant to be fancy, just a fun little experiment to learn something new. Since API credits don’t grow on trees (and neither does CSN), I knew this would be a one-time build. A quick project, just for me, and now, for my portfolio.

The process

First things first: get Spring Boot talking to OpenAI. It took a few cups of coffee and a mild dependency meltdown, but I eventually got Whisper working locally. After that, I spun up a React app with Vite (because fast builds = happy dev). The goal was simple: one page, upload your file, and boom - transcription. I kept the design super clean to make sure the feature stood out, not the fluff.

Problems

Oh, the version drama. I picked Spring AI M2 thinking it would be smooth sailing, but my dependencies clearly had other plans—they wanted to live in the future. That meant cryptic errors, renamed methods, and a bit of confused staring at the screen. Then came the classic “CORS says no” moment when trying to connect frontend to backend. But hey, nothing a custom config can’t fix. In short: a lot of small bumps, but every one was a lesson in disguise.

Lessons learned

• Versions matter: Turns out not all Spring AI work with all versions. Spent an hour wondering why nothing worked... until I checked the docs. Oops.


• Keep it simple: A clean UI doesn't just look nice, it helps people *actually* use your app. Bonus: less CSS-induced headaches.


• Docs are once again your best friend: I used to avoid documentation like a salad at a pizza buffet. Now? It's the first place I go when things break, which is often.