From Zero to Live in Two Days. Vibe Coding a Community Website

I built a full website in less than two days. A couple hours each evening, start to finish. No starter template. No prior experience with Hugo. Just me, VS Code, Copilot, and a vague idea of what I wanted.

The site is for F3 STL County Line, a free, peer-led workout group I’m part of in the St. Louis area. We needed a real web presence: workout locations, schedules, regional links, a guide for workout leaders. The kind of thing that would normally take a week or two of evenings to put together properly.

It took two days.

Starting with a Conversation

I didn’t start by writing code. I started by asking Copilot questions.

What static site generator should I use? Where should I host it? What’s the cheapest option that doesn’t feel cheap? I knew I wanted something fast, easy to maintain, and free, because a volunteer community group isn’t paying for hosting.

Copilot walked me through the options. Hugo for the static site generator. Fast builds, simple content management, good for non-developers to update later. Tailwind CSS for styling. Utility-first, no fighting with CSS specificity. Netlify for hosting. Free tier, automatic deploys from Git, SSL included, custom domain support(I did pay for the domain).

Mostly free. SSL, CDN, continuous deployment, all of it. That part still surprises people when I mention it.

The Build

Once the decisions were made, the actual building was remarkably smooth. I’d describe what I wanted, Copilot would generate it, I’d tweak and move on.

The homepage came together first. A hero section with the F3 mission, cards explaining the three Fs (Fitness, Fellowship, Faith), and the core principles that define every workout. Nothing fancy, but clean and purposeful.

Then the workout locations. F3 calls them AOs (Areas of Operations). Each one needed its own page with a schedule, location linked to Google Maps, the site leader’s info, and a photo gallery. Hugo’s content model made this easy. One template, nine pages of structured content in TOML frontmatter, and Copilot helped me build out the layout for each section.

The Q Card was probably the most content-heavy piece. It’s a comprehensive guide for guys leading workouts for the first time. Warm-up structure, workout flow, the Circle of Trust closing ritual. Copilot helped me organize the layout into color-coded sections that made the information scannable.

Regional pages, responsive navigation with a mobile hamburger menu, footer with Slack invite links. It all came together faster than I expected.

What “Vibe Coding” Actually Means

I’ve seen people use “vibe coding” dismissively, like it means you’re not really building anything. That wasn’t my experience.

What it actually looked like was having a conversation with a tool that could execute. I’d say “I need a responsive grid of workout location cards with images and schedules,” and Copilot would produce something close. I’d adjust the spacing, change a color, restructure a section. It was collaborative in a way that felt natural.

The key was that I knew what I wanted. I’ve been building software for over two decades. I understand layout, responsive design, content hierarchy. I just didn’t know Hugo’s templating syntax or Tailwind’s specific utility classes. Copilot filled those gaps instantly.

That’s the part people miss about AI-assisted development. It doesn’t replace knowing what good looks like. It replaces the time you’d spend reading documentation for a framework you’ll use once.

The Result

A fully responsive community website with nine workout location pages, eight regional links, a detailed leader’s guide, structured schedules, Google Maps integration, and a clean design. Hosted for free on Netlify with SSL and automatic deploys.

Two evenings. Maybe four or five hours total.

A year ago, this would have taken me a week of evenings minimum, mostly spent learning Hugo’s template syntax and fighting with CSS. The actual decisions about structure, content, and design would have been the same. The execution is what collapsed.

---

That’s the real shift with AI tooling. The thinking doesn’t get faster. The typing does.