← Back to all work

Recipe Book: Rebuilding an Early Project With Everything I Know Now

Rebuilding an early coding project with everything I know now: same recipes, same voice, entirely different architecture.

2022 → 2026 rebuilt from scratch
Role
Designer & Developer
Timeline
June 2026
Tools
Astro, VS Code, Claude Code
Team
Solo

The Problem

In 2022, I built a recipe site as an early coding project: HTML files, w3.css, externally hosted images, one page per recipe. It worked. It’s still live. But it was built the way someone learning to code builds things: get it working, ship it, move on. The content was genuine, recipes my partner and I actually cook, with the memories that came with them, but the architecture was whatever got it online that day.

The Process

Browse the recipes →

Early projects are worth revisiting not because they’re bad, but because the gap between then and now is visible. The HTML is fine. The recipes are good. What’s missing is everything underneath: a content model, a design system, a way to add a new recipe without copying and editing another file by hand.

The Decision

Keep everything that made the original worth making. Rebuild everything else.

What stayed:

What changed:

The schema does the same job here that it does for the case studies above: one typed shape that every recipe has to follow, so adding the thirteenth recipe later is a matter of writing a file, not inventing a new pattern.

The Outcome

Reflection

This one’s a reminder that “the gap between then and now” isn’t really about HTML versus a framework. It’s that I didn’t have a way to think about content as a system in 2022, and I do now. The recipes didn’t need fixing. I did.