DocsHow PWARegistry Works

How PWARegistry Works

Given only a domain or URL, PWARegistry:

  1. Fetches the HTTPS page.
  2. Discovers <link rel="manifest">.
  3. Fetches and parses the Web App Manifest.
  4. Resolves start_url, scope, icons, screenshots, and other members against the manifest URL, as the specification requires.
  5. Fetches icon and screenshot files to confirm they are reachable images.
  6. Looks for a service worker registration in HTML and same-origin scripts.
  7. Reads relevant HTML metadata (<title>, description, theme-color, Open Graph) and Schema.org JSON-LD when the manifest omits equivalent fields.
  8. Validates the result.
  9. Normalizes the result with documented rules.
  10. If the result is PASS or PASS WITH WARNINGS and the request was a submission, writes a public registry record.

PWARegistry does not execute JavaScript from submitted sites, does not run a full browser, and does not generate marketing copy.

Standards that are read

Source Used for
Web App Manifest Identity, display, icons, screenshots, capabilities
HTML <link rel="manifest"> Discovery
HTML <title>, meta description, theme-color Fallbacks only
Open Graph og:description, og:site_name Fallbacks only
Schema.org SoftwareApplication / WebApplication Description fallback
navigator.serviceWorker.register Service worker detection

Authoritative how-to documentation lives elsewhere. Prefer MDN, web.dev, the W3C manifest spec, and OpenPWA.net when you need to implement a capability.