Back Donovan Crader Let's talk
Showcase project 2024–2025

ProFuelPrep

A recipe platform built for scale — 204 seeded recipes, AI-generated photography, and a content pipeline under the hood.

  • Next.js
  • TypeScript
  • Express
  • SQLite
  • GPT-4
  • DALL-E 3
ProFuelPrep homepage with a salmon-and-asparagus hero photo, the tagline "Simple, Delicious Meals Anyone Can Make," and a category browser.

Case study

  • Food & recipes
  • Editorial / publishing
  • Content scale

Recipe websites are one of the most saturated content formats on the internet. Standing out requires real photography on every page, fast-loading architecture, and a system that can handle thousands of recipes without becoming a maintenance burden. ProFuelPrep was built to solve all three.

The frontend is a Next.js static export — every recipe renders to flat HTML at build time, so pages load instantly with no runtime failures. A custom recipe index aggregates all 204 recipes, supports ranked search across title, ingredients, and tags, and drives the category browser and time-aware home feed.

The AI content pipeline is the engine behind the site. A GPT-4 script generates a complete recipe — title, ingredients, instructions, prep time, nutrition breakdown — from a single category-and-constraint prompt. A DALL-E 3 step then generates matching food photography at 1792×1024. End-to-end, a production-quality recipe takes about $0.10 and under two minutes. The live showcase runs in read-only mode: visitors browse the full library without triggering API costs.

Content architecture gets as much attention as the recipes themselves. Seasonal landing pages — like "Favorite Fall Recipes for Cozy Season" — collect thematically related recipes with a custom editorial intro. Long-form articles like "Foods That Make The Best Leftovers" provide SEO-targeted content alongside the recipe database. A time-aware home feed surfaces breakfast recipes in the morning and dinner ideas in the evening.

The backend is Express + SQLite deployed on Render, serving from a read-only seeded database in showcase mode. The frontend is a Next.js static export on Vercel. Recipe generation runs as an offline CLI tool — when new content is needed, you run the pipeline locally, commit the JSON output, and redeploy.

ProFuelPrep demonstrates what a serious content-at-scale build looks like: a real AI content pipeline, real food photography on every page, and architecture designed for growth — not a Webflow template with placeholder images.

Screenshots

Like what you see?

Let's build something together.

Have a project in mind? I'd love to hear about it.