Compare a returned route with an official service report

Network Observatory checks its own page host, two fixed public destinations and the official report of the selected OpenAI, Claude or GitHub Copilot service. Requests run directly in the browser when you start a pass; there is no automatic polling or account access.

Open Network Observatory

Keep the four checks separate

The page-host response checks this site's own marker. The public destinations are Cloudflare's documented test edge and GitHub's status host. A readable HTTPS response, including an HTTP error, establishes that this browser reached that tested route. The provider result instead summarizes the selected service's official report.

Choose the next comparison from the result

  1. Read the timestamp and returned state for each check rather than treating the diagram as a discovered router map.
  2. If a public route responded but one app fails, compare that app's official incident scope, sign-in state and extensions separately.
  3. If neither public check is readable, try a controlled comparison with another browser or connection when available; keep the cause open.

Interpret time and browser hints carefully

Displayed durations include the browser fetch and server work. They are not calibrated ping, bandwidth or packet-loss measurements. The online/offline hint can describe a connected local network without working Internet. Official incidents are aggregate reports, not a test of a subscription, account or device.

References used by the tool

Keep the limit in view

CORS restrictions, privacy blockers, endpoint failures and network trouble can all produce an unknown browser read. Two public routes cannot establish the health of the whole Internet or diagnose a router. Providers receive ordinary connection information when the checks are run; the guide receives no test inputs.