SEO and sharing

IWC pages are rendered on the server, so search engines, social networks and AI crawlers read the finished content without running any script. On top of that, IWC writes the technical metadata for you: titles and descriptions, canonical links, robots rules, Open Graph and Twitter tags, structured data, robots.txt, sitemap.xml and correct status codes. None of it adds JavaScript.

Site settings

{
  "name": "Field Notes",
  "head": { "title": "%s · Field Notes", "description": "Essays on making things.", "lang": "en" },
  "seo": {
    "siteUrl": { "env": "SITE_URL" },
    "siteName": "Field Notes",
    "image": "/assets/share.png",
    "twitter": "@fieldnotes",
    "robots": { "disallow": ["/search"] },
    "structuredData": { "@context": "https://schema.org", "@type": "Organization", "name": "Field Notes", "url": "https://notes.example.com" }
  }
}

Per page

{
  "name": "Article",
  "route": "/articles/:slug",
  "data": { "article": { "source": "DATABASE.articles", "mode": "single", "required": true, "filter": { "slug": { "param": "slug" }, "draft": false }, "include": { "author": {} } } },
  "title": "=article.title",
  "description": "=article.summary",
  "meta": { "type": "article", "image": "=article.cover" },
  "structuredData": {
    "@context": "https://schema.org",
    "@type": "Article",
    "headline": "=article.title",
    "datePublished": "=article.publishedAt",
    "dateModified": "=article.updatedAt",
    "author": { "@type": "Person", "name": "=article.author.name" }
  },
  "sections": [{ "type": "text", "variant": "h1", "content": "=article.title" }]
}

The tags are written into <head> and marked so that SPA navigation replaces them when it changes screens. Structured data is written inside a non-executable <script type="application/ld+json">, with <, > and & escaped so content can't end the element.

IWC writes only what you declare, from your real data. It never invents ratings, reviews, prices or authors.

Status codes

• An unknown URL answers 404 with your notFound screen.
• A page whose required page data finds nothing (or nothing the visitor may read) answers 404 too, so /products/no-such-thing is not a soft 404 that search engines index.
• A protected page sends signed-out visitors to sign in (303), and signed-in visitors without the role get 403.

robots.txt

/robots.txt is generated:

User-agent: *
Disallow: /_iwc/
Disallow: /api/
Disallow: /search

Sitemap: https://notes.example.com/sitemap.xml

Protected areas are not listed (robots.txt is public, so listing them would only advertise them). They are noindex instead.

sitemap.xml

/sitemap.xml lists:

• every page without route parameters that a signed-out visitor can open, except the notFound screen, pages marked noindex and pages with "sitemap": false;
• one URL per record for pages with parameters. When a page's data fills each parameter from a field ("filter": {"slug": {"param": "slug"}}), IWC lists the records that page can show, using the same source and literal filters. Otherwise give it a sitemap:

{
  "name": "Event",
  "route": "/events/:year/:slug",
  "sitemap": { "source": "DATABASE.events", "filter": { "public": true }, "params": { "year": "=slice(startsAt, 0, 4)", "slug": "=slug" }, "lastmod": "=updatedAt" },
  "sections": []
}

Records are read as a signed-out visitor would read them: owner rules, roles and tenancy all apply, so the sitemap never lists a record the visitor can't open. lastmod defaults to the record's updatedAt. A sitemap holds at most 50,000 URLs (the protocol's limit for one file); beyond that it is truncated and the server logs a warning.

Semantic HTML

Text variants map to real headings (h1–h6), lists to ul/li, navigation to nav, the page content to main, and links to <a href>, so internal navigation is crawlable even though IWC moves between screens without reloading. iwc check warns about images without alt text and pages with more than one h1.