This started as a brainwave right before bed. I’d been writing about AI-assisted building on LinkedIn, and it hit me: I should track these projects somewhere more structured. More fun. Then the domain idea clicked — one.daybuilds.ai, two.daybuilds.ai… nerds will appreciate how satisfying that is.

What’s this site about?

This is where I document projects built with AI coding assistants—but not just the final code. Each post shares the implementation plan I used to build it.

“Vibe coding” became a thing in early 2025 — Andrej Karpathy’s term for letting AI write your code while you describe what you want. But I use it differently. For me, it’s an accelerator for product managers and hobbyists. A means to an end, not the point. The real skill is in the tools and knowing how to guide them.

Think of this site like open-sourcing the blueprint, not just the building. You can copy these plans, paste them into your own AI assistant (Claude, GitHub Copilot, etc.), and adapt them to your needs.

Why “day builds”?

Because constraints spark creativity. When you know you only have a day (or two), you make different decisions:

  • Simpler architecture - no time for over-engineering
  • Focus on the core idea - strip away the nice-to-haves
  • Ship something - done is better than perfect

Why share implementation plans?

Here’s the thing: your users don’t need your app anymore—they can build it themselves with AI. But they do need good plans.

A good implementation plan breaks the complex into the manageable, captures the decisions that mattered, and doesn’t hide the debugging rabbit holes. That’s what I’m sharing — not just the code, but the thinking that got there.

The duration badges

Each project gets tagged with how long it actually took:

  • 1 Day - Built in a single coding session
  • 2 Days - Needed a bit more time to polish
  • Multi-day - Evolved beyond the initial scope (it happens!)

This site itself

This site? Also a day build. Actually, under an hour — including buying the domains and setting up Cloudflare. Astro made the static generation trivial. Sometimes the constraint isn’t time, it’s just deciding to start.

If even one of these plans saves you a weekend of trial and error, that’s enough.