SEO, structured data and AI discovery

What does the starter publish?

Public SSR pages include a linked JSON-LD graph in the HTML head: the publishing brand, author, website and page. Blog posts add BlogPosting; guides and feature articles add TechArticle. Public inner pages include breadcrumbs. The dedicated FAQ guide derives questions and answers from its displayed content; the template homepage shares one answer list with its FAQ schema.

The template has five launch guides, eight intake groups, two skills, and three payment adapters. These describe repository features, not sales, usage, benchmark results or a promise that providers are configured. See the generation guide, skills and billing.

How do I configure the publisher and author?

Edit identity in app/core/product.ts. Set authorName, authorType (Person or Organization), authorSameAs and brandSameAs. Only add profiles that describe that exact entity. A person’s GitHub profile belongs to the person; a customer site, dependency, competitor or private repository is not a sameAs link for your brand.

Generated products clear the template’s identity links and site publication dates. Set your own facts before launch. Organization identifies the publishing brand; this does not assert incorporation, a registration number or a physical address. Author names in frontmatter matching contentOwner resolve to the configured publisher identity; articles with other bylines retain those names.

Which dates should I use?

Use the real first-publication date and substantive revision date. Content frontmatter supplies publishedAt and updatedAt; the page renders both with HTML time elements and uses them in article schema. identity.publishedAt and identity.updatedAt apply only to the homepage and About page. Do not change dates at every build or backdate a new article.

Do I need special AI markup or growth statistics?

No special markup guarantees AI citations. Google Search Central’s AI features guidance says existing SEO practices apply and structured data should match visible content. Lead useful sections with a direct answer, explain limits, and link named primary sources. Use real product counts or documented measurements; do not invent percentages to satisfy a checker.

FAQPage describes displayed answers; it does not promise a Google rich result. Product, HowTo, ratings, offers and statistics should be added only when the actual page and business support those properties.

How can I verify the result?

Run npm run content:check, npm run typecheck, npm test and npm run build. Inspect View Source: JSON-LD should be in the head and readable without browser JavaScript. Check the canonical, author links, dates, Open Graph and Twitter card image. Drafts, account pages and internal search remain noindex and omit the public entity graph.

Use Google’s structured data testing guidance and article documentation for external checks. A local pass or valid markup does not prove indexing, a knowledge panel or inclusion in an AI answer.