I Slipped and Fell on an App Store Listing
- Rich Washburn
- 7 hours ago
- 5 min read
I want to be clear about something upfront: I claim zero talent or skill in what I'm about to describe. This is a broken clock story. The clock just happened to be right at a useful moment.
Labor Day weekend 2024. Raining in South Florida. Bored enough to do something stupid. I grabbed a Xiao ESP32-C3 — tiny little microcontroller, about $5, probably the smallest ESP32 package I've ever seen — and eight buttons. Wired four of them to standard shortcuts: copy, paste, cut, enter. The other four got something more interesting. Each one was programmed to spit out a block of preset text the moment you pressed it. Long text. Fast. Like pasting, but in hardware.
Prompts, specifically. We were deep in the GPT-3.5/4.0 era and I was tired of typing the same context into seventeen different sessions. So I built a thing that did it for me. One button. Done. I called it the Prompt Potato Gun. Because of course I did. The thing had Bluetooth, a config mode over Wi-Fi where you could swap out your "potatoes" without reflashing anything, RGB LEDs for status (and honestly just because they look cool), and eventually a custom PCB I designed in EasyEDA and had manufactured at JLCPCB. I built about a dozen of them over the following months. Gave most of them away. The whole first one cost five dollars. The rainy day was free.
The idea never fully left me. I just wanted a software version of it. Something that lived above my other windows, stayed small, and let me grab a chunk of text and drop it wherever I needed it — without hunting through eleven browser tabs or the Notes app graveyard I've been meaning to clean up since approximately 2022. So a few weeks ago I'm killing time between actual work and I find this Mac app builder called Glaze. They hand you 120 free tokens to start. I figured, what the hell.
I asked ChatGPT to sketch a design concept. Dropped it into the Glaze prompt. Hit plan. Watched it think for a couple of minutes. Hit "do the plan." What came back was 90% of PromptPocket.
I wish I could tell you there was a vision here. A roadmap. A product spec.
There wasn't. What happened was: it worked, and then I started adding things. Tabs. Nodes inside the tabs — each one its own little block of text. Auto-detection for URLs so links get a Launch button instead of just sitting there. And then the one that actually changed how I use the laptop at night: voice dictation on every node. Being able to lie in bed, tap a button, talk a thought into a node, and then grab it in the morning Small win. Completely real win.
I added an About card. Click the coffee cup at the top, it goes straight to my site. Because if I'm accidentally shipping software, I'm at least making it point back to me. Total tokens burned for all of that, including voice dictation: I still have 30-something left out of my original 120.
Here's the thing about vibe coding that nobody really wants to say out loud. Most of the time it's a slog. You spend hours watching the AI confidently do the wrong thing, rebuilding what it just broke, and eventually end up with something that technically runs but you'd never actually use. It's fun. It's genuinely interesting. It scratches the builder itch. But "load-bearing" and "enterprise grade" are not words that usually apply to the output.
PromptPocket is wildly stable. That surprised me more than anything.
I published it to the Glaze store — first time I've had anything listed in an app store, which is its own weird little milestone — and started passing it around. The response was not what I expected.
A sysadmin friend of mine, runs IT infrastructure for a fairly large company, looked at it for about thirty seconds. Shook his head. Said: "Why does this not already exist? Why is this the first time I've seen this?"
I didn't have a great answer. It's a simple idea. It should have existed. Apparently it didn't, at least not in a form that fit the workflow. So now it does.
Then I hit the obvious wall. Most of the people I thought would actually get the most use out of this — people hammering LLMs all day, people who copy the same system prompt into a new session for the eighth time this week — they're on Windows. The Mac native app, running on the Glaze runtime, doesn't help them. So I rebuilt it.
The web version is a Progressive Web App. Any modern browser. Chrome, Safari, Edge — doesn't matter. You install it from the browser, it sits on your desktop or taskbar like a native app. You lose the floating-above-everything behavior of the Mac version, but the workflow holds.
Tried it on iPhone. Works. Hit share in Safari, save to home screen, launches like an app. Haven't tried it on Android yet but I'd bet it works.
It has its own domain: promptpocket.space I paid for the year. So it'll be free for at least that long. You're welcome.
The Mac version needs the Glaze app installed alongside it — that's the runtime, it's free, clean install, nothing weird going on. They give you 120 tokens to start building your own stuff, which is genuinely enough to ship something useful if you don't iterate yourself into a corner the way I tend to. The web version needs nothing. Open the URL. Install if you want. Use it anywhere. Both are free.
The Prompt Potato Gun was hardware because in 2024, software was still too slow and awkward to deliver this in a form I'd actually use. PromptPocket is software because the tools finally caught up to the idea. The idea was always the same. The constraint just changed. That's what I keep thinking about. The things that were too annoying to build two years ago — too small, too personal, not worth the overhead — they're getting their moment now. Not because the ideas got better. Because the cost of attempting them collapsed.
PromptPocket passed the only test that actually matters: does this make something slightly less annoying? Yes. Immediately. For me, and for a sysadmin in Connecticut who manages infrastructure for hundreds of people and still had the same clipboard problem I had on a rainy Sunday in Florida. That's enough.
Get it: - Mac (native, via Glaze)
More info: https://www.richwashburn.com/ppocket
Rich Washburn is a technologist and strategist working at the intersection of AI, infrastructure, and capital. He is Managing Partner and Chief AI Officer at Eliakim Capital and CIO of Data Power Supply.
Comments