The Evolution of AppBunker — From a Notion Database to an AI-Powered Tool

A Seed Planted Years Ago

Several years ago, I launched a modest side project on Gumroad — a resource library housed in a Notion database (check it out here: https://iamjesushusbands.gumroad.com/l/resource-lib). It was a personal endeavor, a catalog of hundreds of applications I’d stumbled upon over time — tools ranging from productivity boosters and development frameworks to creative utilities and niche solutions. I offered it for free, expecting little beyond my own use. What unfolded surprised me: users found it valuable enough to contribute voluntarily, turning a simple database into a small but revenue stream. That organic support hinted at its potential, but for years, it remained a quiet repository, growing incrementally as I added new finds.

Fast forward to today, April 11, 2025. That humble Notion database has evolved into something far more dynamic: AppBunker (https://app-bunker-melior.replit.app/). What began as a static collection is now an AI-powered side project, fueled by a large language model (LLM) and web search APIs — and remarkably, it came together in just a few hours thanks to a handful of vibe-coding prompts. This blog post traces that journey, from its roots to its current form, and outlines why I built it: to streamline the discovery of tools that enhance my work and solve real problems for my clients.

The Original Vision: A Library of Possibilities

The Gumroad library wasn’t born from grand ambition — it was a practical solution to a personal pain point. As a software developer running an agency and juggling subsidiaries, I constantly encountered tools that promised efficiency, innovation, or just a better way to get things done. Some were game-changers; others fizzled out. Tracking them became a chore — bookmarks got messy, notes scattered — so I turned to Notion. I built a database, organised by use case, platform, and value: project management apps, prototyping suites, analytics dashboards, you name it. Hundreds of entries piled up over the years, each a snapshot of my tech explorations.

I listed it on Gumroad with a $0 price tag, thinking others might peek at it. To my astonishment, people didn’t just download it — they chipped in. Contributions trickled in — $5 here, $10 there — proof that this unpolished collection resonated beyond my own desk. It wasn’t about the money (though it bought a few snacks); it was the signal that others saw utility in what I’d curated. That validation planted a seed, but life kept me busy — client projects, tech trends, the grind. The library sat, quietly useful, until a recent spark reignited it.

The Turning Point: Vibe Coding Meets Opportunity

Intuitive approach to building with AI tools like Replit’s Agent and ChatGPT. I’ve been deep in this space, prototyping apps in hours, not weeks, and pushing the limits of what these tools can do. One day, tinkering with a client spec, I realized my old Notion library could be more than a reference — it could act. What if I layered intelligence on it? What if it didn’t just list tools but helped me find them, pulling from its own data and the web?

I started with a simple goal: transform this static database into a dynamic search tool. I turned to ChatGPT, feeding it the library’s structure and my vision: a system to query apps by need — say, “prototyping tools for mobile” or “analytics for small teams” — and get results fast. It spit out a set of vibe-coding prompts, concise but sharp, tailored for Replit’s Agent. I plugged them in, hit run, and within a few hours — yes, hours — AppBunker emerged. An LLM parsed the database, APIs tapped web data, and a rough-but-working interface tied it together. It wasn’t just a list anymore; it was alive.

AppBunker: What It Is and Why It Matters

AppBunker isn’t a polished product yet — it’s a side project with promise. At its core, it’s a tool discovery engine. The original library’s hundreds of apps form its foundation, but the LLM lets me ask natural questions — “What’s good for automating workflows?” — and get tailored answers. The APIs extend its reach, scraping real-time info from the web to fill gaps or update stale entries. In one test, I queried “client dashboard tools,” and it pulled options from the database (like a CRM I’d logged in 2022) while snagging fresh alternatives online. Rough edges aside, it worked.

My purpose is pragmatic: streamline how I find tools for myself and my clients. Running a dev agency means constant problem-solving — whether it’s a quick prototype for a pitch, analytics for a startup, or operations tweaks for efficiency. AppBunker cuts the hunt, saving hours I’d spend googling or scrolling Product Hunt. It’s personal, too — my years of curation now serve a sharper purpose. But it’s not just for me. I see it as a resource others could tap, especially those navigating the same tech overwhelm I’ve wrestled with.

The vibe-coding angle is what floors me. A few prompts — exactly 10 — turned a dormant database into this. No weeks of coding, no team, just me, Replit, and AI. It’s a testament to how far these tools have come — ChatGPT shaped the logic, Replit’s Agent wired it up, and APIs breathed life into it. Hours, not months. That speed’s reshaping how I think about building.

The Bigger Picture: From Side Project to Potential

AppBunker’s early, but its potential’s brewing. For me, it’s a daily driver — already helping me pick tools for client gigs and my own experiments (ML Ops, anyone?). I’ve always believed in solving real problems, and this one is a start: too many tools, too little time. But I’m not blind to its limits — UI’s basic, search can hiccup, and scaling it needs work. That’s where the journey’s at: refining it into something broader, maybe even a public resource.

The Gumroad library’s community support years ago showed me people crave curated solutions. AppBunker could take that further — imagine a platform where devs, founders, or even non-techies query their needs and get answers, backed by AI and a growing database. It’s not there yet, but the bones are solid. I’m committed to tweaking it, and that’s where input matters. I’ve linked both the original Notion library and AppBunker above — explore them, poke around, tell me what’s missing. What features would make this a go-to for you?

Closing Thoughts: A Work in Progress

This evolution — from a free Notion database to an AI-powered tool — blends years of curation with today’s tech. AppBunker’s not a finished product; it’s a proof of concept, born from vibe coding and a desire to work smarter. It reflects my ethos: build what solves, share what helps. The contributions from that Gumroad launch years back lit a spark; this is me fanning it into something new.

I’m eager to hear your take. Try AppBunker, revisit the library, let me know what clicks — or doesn’t. What tools do you lean on? What gaps do you see? This is an ongoing effort, and your insights will shape its next steps. For now, it’s a side hustle with legs — and I’m excited to see where it runs.