ReligionCompare

Overview

ReligionCompare is a comparison site covering roughly two dozen religious traditions across several hundred pages. It explains beliefs, practices, histories, and the similarities and differences between traditions, in a structured and neutral way. It is the portfolio’s clearest example of a content site built around a single non-negotiable principle: trust.

The goal

A religion comparison site fails the moment a reader does not believe it. So the make-or-break feature was never the design or the page count. It was citations. Every substantive claim needed a source behind it. The goal was a site that earned trust by construction, not by tone.

The stack

Next.js, deployed on Cloudflare with D1 for the database, R2 for storage, and KV. The content architecture matters more than the framework here: a religion taxonomy, a denomination taxonomy, a D1 schema built specifically to attach sources to claims, and admin workflows for managing structured, citation-backed content rather than freeform articles.

What I built

The comparison engine and the taxonomy underneath it. Structured content where traditions and denominations are first-class data, not just pages, which is what makes real comparison possible instead of a pile of disconnected articles.

The citation system is the heart of it. Claims are tied to sources at the data level, and the admin workflow is built around that model rather than around a generic blog editor.

On top of that, SEO and technical architecture at scale. Several hundred pages means programmatic content done carefully. Canonical-host handling. Structured data. Sitemap discipline. Locale routes deliberately blocked from crawling so they could not fragment ranking signal or create duplicate-content problems.

The hard parts

The honest version. Programmatic content at this scale is a constant fight against thinness and duplication, and a content architecture that does not treat that as a first-class problem will produce hundreds of pages search ignores. The hardest single category of bug was infrastructure rather than content: canonical-host and redirect logic behind a proxy. Middleware written as if the app receives traffic directly will compare against the wrong host and protocol values once it sits behind Cloudflare, and the result was a redirect loop that took real debugging to unwind. The fix was using the forwarded-host header the proxy actually provides, and the broader lesson, that you verify against the live domain and not just a successful build, came straight out of that work.

Where it landed

ReligionCompare stands as a several-hundred-page citation-backed reference site, with a content architecture built for trust and for search at the same time. It is the project that proved a content site does not have to choose between credibility and scale, as long as the architecture treats both as requirements from the start.

My role

Solo. Content architecture, taxonomy design, build, and SEO.