The Great Agnostic — a literary archive in monochrome

The brief, framed honestly

The Great Agnostic is a personal project: a literary archive of Robert Green Ingersoll, the nineteenth-century orator known to history as "the Great Agnostic." His complete works fill twelve volumes in the Dresden Edition, and at the time of this build, no clean, reader-friendly, search-engine-indexable version of the full corpus existed online. Plain-text scans on archive.org. PDFs on Project Gutenberg. A few of the more famous lectures floated around in blog excerpts. Nothing approaching a real archive.

This is the entire archive, on the open web, with a structure that respects both the reader and the search engine. Built greenfield, no pre-existing site to migrate, no CMS to fight with.

The build, end to end

The full Dresden Edition was ingested from source TXTs into Eleventy via a custom pipeline (scripts/ingest-dresden.mjs) that reads a manifest of WORKS, identifies each work's body content by a distinctive start-marker phrase, and emits per-work markdown files into the Eleventy collection. 177 works in total, each with its own URL, breadcrumb, structured-data block, and prev/next navigation.

The visual register is Victorian. Playfair Display for display, Crimson Pro for body, monochrome palette throughout (paper warm-cream, ink dark charcoal, no chromatic accent). The typographic system carries drop caps on long-form work, italic emphasis preserved from the source text, and a sticky table of contents on every long piece.

Beyond the works themselves, the site ships a full biography, a 25-event timeline, a connections graph linking Ingersoll to thirteen contemporaries (Whitman, Twain, Lincoln, etc.), a resources page with the original Dresden Edition source links, an FAQ, and a privacy/terms layer for completeness.

Search runs client-side via Pagefind across every word of every work. RSS feed for the blog. JSON-LD on every page in the right shape (CreativeWork for works, Person for the biography, BlogPosting for posts, FAQPage for the FAQ).

The principles applied

Custom code over CMSes

Eleventy v3 (ESM), Nunjucks templates, vanilla CSS, vanilla JS. No WordPress, no headless CMS. The 177-work corpus would have needed its own migration tooling on any CMS; on Eleventy each work is a markdown file, and the build emits the entire archive in under three seconds.

Fast by construction

Mobile PageSpeed 95–100 across every page, including the works themselves which run several thousand words each. CSS concatenated and per-page-purged at build time. No analytics tracker. No third-party fonts on the visitor path.

Accessible by default

Long-form Victorian prose has its own accessibility hazards — drop caps, all-caps section heads, italic emphasis dense enough to fight contrast. Each was solved by hand: drop caps use real :first-letter rules with adequate size and color, all-caps is reserved for section labels not body, italics retain full text contrast. WCAG 2.2 AA across the corpus.

Boring infrastructure

Eleventy + Nunjucks + Cloudflare Pages. No build step that depends on a vendor's account, no deploy that depends on a runtime service. The site is files in a repo, hashed and deployed.

Why it qualifies as a case study

Most service-business websites are a tenth this complex: ten or twenty pages, a blog, a contact form. The Great Agnostic is 177+ pages of long-form historical prose, custom-typeset, fully structured for search, with a working theme toggle and a working search and a working RSS feed and a working FAQPage rich-result. If the engineering posture can produce that, a roofing site or a plumbing site is a smaller version of the same job.

Visit the live site

The whole archive is on the open web at thegreatagnostic.com. Run any URL through PageSpeed Insights. View source on any work. The page-speed numbers, the structured-data shape, and the accessibility posture are all auditable in the public tools.