Many Claude users could not log in to claude.ai or Claude Code on the morning of November 13, 2026. Anthropic reported the problem at 09:58 UTC (4:58 a.m. ET), deployed a fix at 11:05 UTC and marked it resolved at 11:58 UTC (6:58 a.m. ET), about two hours in total. The Claude API, used by developers and apps, stayed up. A separate problem with Claude’s Microsoft 365 connector, caused by a major Microsoft outage, was still unresolved at Anthropic’s last update.
What happened to Claude this morning?
Anthropic’s status page opened an incident titled “Elevated errors on claude.ai, Claude Code, and platform.claude.com,” saying: “We are investigating an issue with users authenticating to Claude.ai and Claude Code.” Authenticating simply means logging in. People who were already signed in could sometimes keep working, but new logins failed. Anthropic has not published a cause.
Which Claude services were affected?
During the incident, the status page showed:
- claude.ai: major outage
- Claude Console (platform.claude.com): partial outage
- Claude Code: partial outage
- Claude Cowork: degraded performance
- Claude API: operational, so apps calling Claude directly mostly kept working
Is the Microsoft 365 connector still broken?
It was at Anthropic’s last update. Since 17:52 UTC on November 12, Claude’s Microsoft 365 connector has failed for most mail, calendar and search requests, and Claude for Outlook has been affected, with some users unable to open the add-in. Anthropic says the cause is a Microsoft 365 service incident, not Claude itself. Microsoft’s outage hit Outlook, Teams, SharePoint and Copilot worldwide, with more than 130,000 Downdetector reports for Outlook alone, Startup Fortune reported. Microsoft blamed infrastructure problems and said there were no signs of malicious activity.
Has Claude had other outages recently?
Yes, though mostly short ones. The status page also lists brief model errors on November 8 and November 10, and a Google Calendar connector problem from October 26 to 28. The longest recent disruption we covered was the September 29 outage. The timing is awkward: Anthropic is days away from its stock market debut; see Anthropic’s IPO plans.
What should I do if I cannot log in to Claude?
- Check status.claude.com first, before reinstalling apps or extensions
- Wait and retry once the status shows recovery; clearing cache rarely helps during a server-side incident
- If you see a security block, try a different network
- Do not reset passwords or rotate API keys just because login failed; only do that if a secret was exposed
- Ignore “recovery” links in emails or messages; scammers use outages to phish
How can teams protect themselves from AI outages?
Keep a fallback. Developers can route important jobs through the API, which stayed up today, or keep a second model provider ready; see our model comparison. Teams relying on Claude Code should pause merges that depend on it until service is back, and avoid building critical workflows on a single tool. Our guide to the best AI coding agents covers alternatives.




