google search index

docs/design/website-indexing.md

@@ -0,0 +1,47 @@
+# Website indexing (Google)
+
+## User + job
+
+Someone searching for a BYO-LLM writing/code assist tool finds LexAI and lands on the matching page (home, docs, contribute, or community). Success: the snippet title and URL match the page they open.
+
+## Flow
+
+Search result → HTTPS path (`/`, `/docs`, `/contribute`, `/community`) → page. Old `#/docs` links redirect to `/docs` with no extra click. Unknown paths: one screen, link home. Target: 0 taps after the SERP click.
+
+## Screen states
+
+| State | What |
+| --- | --- |
+| Success | Existing Home / Docs / Contribute / Community; browser tab title matches the route |
+| Not found | H1 “Page not found”, one sentence, primary “Home”, secondary “Docs” |
+| Loading / empty / offline | Unchanged marketing pages (static). No new loading chrome |
+
+## Components
+
+Reuse `SiteHeader`, `SiteFooter`, `.page` / `.docs-content` / `.btn`. No new design-system pattern except **Not found** (same as `.page`).
+
+## Copy (en)
+
+- Title (404): Page not found
+- Body: That address is not a LexAI page. Check the URL or go back home.
+- Primary: Home
+- Secondary: Docs
+
+Route titles/descriptions live in `src/seo.ts` (not on-page H1s).
+
+## Accessibility
+
+404 is a real `h1`; both actions are text links/buttons ≥ 40px; focus order follows visual order.
+
+## Acceptance
+
+1. `/docs`, `/contribute`, `/community` are path URLs (not `#/…`).
+2. `/robots.txt` allows `/` and names the sitemap.
+3. `/sitemap.xml` lists the four indexable URLs on `https://lexai.juankibin.space`.
+4. Each of those URLs has a unique `<title>` and meta description in the HTML file Google fetches.
+5. Unknown path shows the 404 copy; crawlers get HTTP 404.
+6. `#/docs` becomes `/docs`.
+
+## Out of scope
+
+Search Console verification token, sitemap ping, paid ads, i18n locales beyond en.