Building My Own Slip-Box (with Audio and AI)

Building the Zettelkasten Virtual Voice Box

By Jed May ·

The past few days I’ve been vibe-coding a little side project, my first time ever vibe-coding anything, actually. I don’t love the style, but I needed the experience. I’ve contributed heavily to a MAUI app before and know the pieces can work together, but this is the first time I’ve tried pulling it all up from scratch. And honestly, it’s been a decent taking-off point.

The idea started when I read How to Take Smart Notes. That book kicked everything off for me, and I’ve been hooked on the concept ever since, even pulling Obsidian into my daily routine.

So what is Zettelkasten (aka the Slip-Box Method)? In short: you write each note like its own self-contained thought. One slip = one idea. Later, you link those slips together. The point isn’t just to store information once and forget it — it’s to build a living web of ideas that keeps sparking new ones. The magic isn’t in a single note, it’s in the connections between them.

Later, I stumbled onto How I Built My Second Brain in 3 Hours by Jonas Braadbaart from The Circuit. His project used voice to record notes, with workflows automatically categorizing everything. That clicked with me. I even tried putting my own spin on it with Tasker on Android. While it technically worked, it felt clunky and way too much overhead.

That’s when I started wondering if I could hack together my own version.

Most of my creative thinking happens (against my will) when I’m trying to go to sleep. I thought: how amazing would it be to just hit record on a widget, capture a thought, and have it actually be useful later? Not just categorized, but following that slip-box approach. Perfect for summarizing chapters of a book, jotting thoughts on a blog post, or catching those random 2 a.m. ideas before they vanish.

Here’s the flow I’ve been working on:

  • I record a thought — a book summary, a reaction to a post, or some random late-night idea.

  • OpenAI’s transcription endpoint spits back a verbatim transcript. No cleanup. The messiness stays, along with the original audio file.

  • I tag it with metadata either before recording or after saving if it’s a book (title, chapter, page, etc.), or if it’s just a random idea/thought, leave it alone. Then stash it away.

  • Everything lives locally as JSON + audio files. No database. Just my own slip-box, stored on my device (maybe synced to Google Drive later).

Here’s the magic: when I want to write about something, I type in a general topic. Crazy example, let’s say, “how jellyfish are immortal.” The app sends that to OpenAI, which digs through all my notes. Not just for keywords, but for weighted ideas. It might pull back notes related to fish, jellyfish, immortality, morality — whatever’s even loosely connected. Then it distills a consensus out of those notes into a sentence or paragraph.

That’s where it gets interesting: I could land on an original idea I didn’t even realize was hiding in my own words.

Will anyone else ever use this? Probably not. It’s a pet project. If I ever shine it up, maybe I’ll toss it on the Play Store, but that’s not really the point. It’s teaching me a lot about MAUI, about how I think, and about how raw thoughts — when captured consistently — can snowball into something bigger. Most importantly, it’s making me think hard about user experience: how effortless can I make it for someone to capture a thought and let the slip-box method do the rest?

At the end of the day, this thing might never go beyond my phone, and possibly a few others who have expressed some genuine interest in just the recording audio and transcribing functions. But as a playground for learning and experimenting, it’s already worth it. And if nothing else, I’ve proven to myself that vibe-coding doesn’t always mean wasted coding.

An unhandled error has occurred. Reload 🗙