@burdenoff/website-sdk 2026.1021.2 → 2026.1021.4

package/README.md

@@ -318,6 +318,64 @@ attributes are deliberately distinct: several assertions select
 back, so a second element wearing it would keep them green while describing the
 wrong element.
 
+### Legal pages (`LegalPage`)
+
+Every product website publishes the same legal set, read from the global CMS.
+A product either has its own row for a slug or falls back to the global
+template, whose `{{productName}}`-style placeholders are filled in by
+`publicResolvedContentPage` (always use the resolved query; the plain one
+returns raw placeholders for the templates).
+
+| `document` | CMS slug |
+|---|---|
+| `privacy` | `privacy-policy` |
+| `terms` | `terms-of-service` |
+| `cookies` | `cookie-policy` |
+| `refunds` | `refund-policy` |
+| `aup` | `acceptable-use-policy` |
+| `dpa` | `data-processing-addendum` |
+| `ai-terms` | `ai-product-terms` |
+| `subprocessors` | `subprocessors` |
+
+For prerendered sites, fetch at build time and pass the result in, so the page
+renders its final HTML with no request and no layout shift:
+
+```ts
+// scripts/fetch-legal-content.ts
+import { WebSDKClient } from '@burdenoff/website-sdk/client'
+import { fetchLegalContent } from '@burdenoff/website-sdk/content'
+
+const client = new WebSDKClient({
+  gatewayUrl: 'https://graphql.burdenoff.com/global/graphql',
+  productId: 'vibecontrols',
+})
+// Rejects on any failed request: keep the previous JSON instead of writing an outage.
+const pages = await fetchLegalContent(client, { productId: 'vibecontrols' })
+await Bun.write('src/data/legal-content.json', JSON.stringify(pages, null, 2) + '\n')
+```
+
+```tsx
+import { LegalPage } from '@burdenoff/website-sdk'
+import legal from '@/data/legal-content.json'
+
+<Route
+  path="/dpa"
+  element={
+    <LegalPage
+      document="dpa"
+      initialContent={legal['data-processing-addendum']}
+      productName="VibeControls"
+      contactEmail="legal@vibecontrols.com"
+    />
+  }
+/>
+```
+
+Without `initialContent` the page fetches at run time through the
+`WebSDKProvider` client. A document the CMS does not have renders an honest
+"not available" state marked `noindex`. A leading `# Title` in markdown content
+is lifted into the page heading, so there is only ever one `<h1>`.
+
 ### SEO Component
 
 ```tsx