AI App Development

LunyaCatcher.ai

An AI Speech-to-Text App That Respects Your Privacy

TL;DR — LunyaCatcher.ai is a privacy-first, local voice-to-text app built in under 24 hours using Python and Flask. It leverages OpenAI's Whisper model to transcribe speech without sending data to the cloud — perfect for creatives and small teams who value control over their ideas. The project showcases how AI can assist with branding, design, and development under real-world time constraints.

View Demo

🧩 The Problem

I needed a speech-to-text tool I could trust — but every app I tried came with a caveat: cloud processing, data logging, or some level of privacy compromise. I wanted to create, but not at the cost of feeding my voice to a tech giant.

đź’ˇ The Solution

Meet LunyaCatcher.ai: a local voice-to-text app that lives on your computer — and only your computer. No big tech. No cloud processing. Just your voice, your data, and your creativity.

đź“– Backstory

In the summer of 2025, I started testing AI voice tools during my 3-hour daily commute. Some apps were impressive, but none fully earned my trust. I knew most were training their models with my data — and I didn’t want to be part of that.

So I gave myself a gift: one full day to build something just for me. That’s when the idea for LunyaCatcher took shape. I needed something private, lightweight, and functional.

Initially, I considered using the JavaScript API — but even that wasn’t fully private. Then I discovered OpenAI’s Whisper model, which supports local transcription with powerful accuracy.

" Whisper is an open-source speech recognition system by OpenAI that converts spoken language into text. It runs locally, supports multiple languages, and works without needing an internet connection."
— ChatGPT

🛠️ The Build: Tech Stack

  • Python

  • Flask + Jinja

  • HTML + CSS

  • JavaScript

🎬 Demo Time

Want to see it in action?

Here’s a quick overview of how the app works:

  • Navigate to the upload page.

  • Confirm your upload and submit the file.

  • Transcription begins automatically after submission.

  • Review the transcribed text carefully.

  • Copy the text into your preferred editor to continue working.

🎨 Branding & Design with AI

As a full-time mom managing a busy household, I don’t always have long hours to spare — so when I gave myself one day to build this app, I had to be intentional with every decision. I wanted to focus on coding and problem-solving while still delivering a product that felt cohesive and well-designed. That led me to explore how AI could support the creative side of the process without taking over.

  • I generated the app’s original logo with AI, based on a moodboard I created.

  • I used AI to draft presentation slides for the About page, helping me communicate the app’s purpose clearly (See demo below.)

  • I translated hand-drawn sketches into Figma-ready assets, streamlining the transition from idea to UI.

AI didn’t replace the design work — it accelerated it. I stayed in control of the direction, using AI as a tool to bring the vision to life faster and more efficiently.

👤 Who It’s For

LunyaCatcher is ideal for:

  • Writers, creatives, and thinkers who want their words in text form — without sacrificing privacy.

  • Small business owners who need on-device transcription without relying on cloud tools.

  • Anyone who wants to move away from the keyboard and get their thoughts down with their voice.

⚙️ Challenges & Wins

Building this in just one day was a real challenge. I had to stay focused on what mattered most: making something useful, private, and easy to use. Getting back into Python after some time was a bit of a stretch, but it also reminded me how much I enjoy working with it. One of the biggest wins was finding ways to use AI without compromising my values. I made sure the app stayed local and respectful of users’ privacy — exactly what I wanted from the start.

🔮 What’s Next?

Here’s what I’d like to add in future versions:

  • In-app voice recording

  • Login/user profiles

  • Built-in text editor

  • Transcription formats (PDF, DOCX, etc.)

  • Summary generation & content repackaging

  • Multilingual transcription support

  • Local save and clip management

✨ Why This Matters

This project isn’t just a cool build — it’s a reflection of how I think about tech: ethical, useful, human-first. I created LunyaCatcher for people like me — people who want to build and create without compromise.

đź”— Explore More

If you enjoyed this case study, check out my other projects below.

Other Projects