JSON-LD for Personal Sites: Structured Data That Boosts SEO and LLM Citations

JSON-LD lets you embed machine-readable structured data in a webpage via a <script type="application/ld+json"> block that browsers ignore but crawlers like Googlebot parse. The data follows the Schema.org vocabulary and is modeled as a labeled directed graph, where each node carries a @type (e.g. WebSite, WebPage, Person), a unique @id (typically a URL plus a hash fragment), and descriptive properties. Crawlers can merge node properties across pages when IDs match, so the author recommends fully detailing core nodes on the root page and shipping slimmed-down versions elsewhere—though single-page scrapers like LLMs won’t perform that merge.

The practical payoff is richer link previews, clearer site labeling in search results, and better discoverability. The author walks through the key node types for a personal site: WebSite for site-level metadata, WebPage (and subtypes like ProfilePage and CollectionPage) for the actual HTML page, and Person to describe who you are. The Person node is the most consequential—Google folds it into content-quality signals, and the sameAs array (linking GitHub, LinkedIn, HN, and other profiles) helps crawlers disambiguate common names and assemble a knowledge-graph identity.

The notable wrinkle is the AI angle: the author observes that LLM crawlers increasingly read Person data to decide whom to cite in generated answers, making structured data relevant not just to traditional SEO but to visibility in AI-mediated search. The post doubles as a copy-paste template, with the author’s own JSON-LD provided for each node type.