Check whether Grok is down before you troubleshoot

“Something went wrong” on grok.com and a failed reply in the Android app can share one cause that no reinstall touches. status.x.ai reports each surface on its own row: Grok (iOS), Grok (Android), Grok (Web), API (us-east-1.api.x.ai), and API (us-west-2.api.x.ai). On Sep 3, 2026 that page carried a Models outage banner with Web, Android, and both API regions in outage while iOS stayed available.

Read the row that matches your surface

1. Open status.x.ai and find the row for where you are typing.
2. Scroll to Live service data, which is exported from xAI monitoring and shows service health even when no incident has been declared.
3. For the API, check the endpoint you actually call — eu-west-1, us-east-1, and us-west-2 are reported separately, so one region can serve while another fails.

An outage row explains repeated failures across accounts, browsers, and devices. Waiting is the fix, and switching to a surface still marked available keeps you working.

When the status page is clean

Official troubleshooting for account and app failures starts with restarting the app, deleting and reinstalling it, signing out, and clearing the cache. On the web, Chrome or Edge in an incognito window isolates browser-specific breakage. Then use Report an issue in the product with your account email, the platform, your browser and OS, and a screenshot or conversation share link.

Grok in X and Grok Bot

Grok inside X runs on X's service, and xAI says it has no operational oversight there — those reports go to the X Help Center or @premium on X. Grok Bot failures follow a separate path, since a stuck Agent Computer is not one of the rows on this page: Fix Grok Bot “Can’t reach your computer”.

Pitfalls

• A green board plus a broken account usually points at billing or sign-in state.
• One region recovering does not clear the others.
• Repeated reinstalls during a declared outage cost you the sign-in, with no gain.