Yurasis Article

The article reader lives at /[lang]/article/[blogSlug]/[postSlug]. The blogSlug segment selects which blog's configuration to apply — primary color, name, categories, AdSense publisher ID, and site verification tags all come from the blog record. There is no per-blog deployment, no per-blog codebase, and no per-blog build.

A special slug value ("portfolio") short-circuits the lookup and renders portfolio case study content instead of a database post — the same route handler serves both article reading and the portfolio case study pages without a separate route.

Each article page injects BlogPosting (with headline, description, author, publisher, image, wordCount, and timeRequired), Blog (with publisher reference), BreadcrumbList (three-item chain: Yurasis → blog → post), and FAQPage (when the post stores a faqSchema array). All four blocks are rendered as <script type="application/ld+json"> in the server HTML.

hreflang alternates are derived from the post's translationOf and translations fields. canonicalTranslation() picks the authoritative locale-slug pair, preferring the requested language, falling back to the blog's defaultLocale, then ko/en. The canonical URL is always the default-locale version of the post.

When a post has a category, the reader fetches up to six posts from the same category, excludes the current post, and renders the first four as a related posts grid below the article. When no category exists the query falls back to the post's first tag. No embeddings, no ML model — just a filtered API call at request time.

The related post cards link to the canonical path for each post, respecting the language preference of the reader. If a translation exists for the reader's locale, the card links to that translation.

View count is incremented by a ViewTracker client component that fires a POST request after hydration using a keepalive fetch. The server renders the article without waiting for the view event — the count the reader sees is viewCount + 1, accounting for the current visit before the API confirms it.

Daily metrics are aggregated separately by a cron job that writes to blog_daily_metrics, keeping the hot path (article render) free from analytics writes. The dashboard reads from the aggregated table, not the live viewCount column.