Anthropic fixes authentication errors that disrupted claude.ai, Claude Code and Claude Console

Anthropic's Claude platform suffered a bout of elevated errors on November 13, 2026, after a problem with user authentication spread across the company's consumer and developer products. According to the incident report on Anthropic's status page, the fault affected claude.ai, Claude Code, Claude Console and Claude Cowork, and a fix was deployed a little over an hour after the issue was first acknowledged.

What happened

The first public sign of trouble came at 09:58 UTC on November 13, when the Claude status page was updated to say Anthropic was looking into an issue with users authenticating to Claude.ai and Claude Code, and that an update would follow as soon as possible.

At 11:05 UTC, one hour and seven minutes later, the page was updated again to report that a fix had been deployed for the problem affecting Claude services and that engineers were monitoring to confirm recovery. As of that last update, the incident remained in a monitoring state rather than being declared fully resolved, and neither update explained what had actually caused the authentication failure.

The stated symptom is notable: the initial update pointed at sign-in rather than model availability. In other words, the models may have been running, but users could not reliably get to them.

Which products were hit

The incident report lists four affected products:

• claude.ai, the consumer chat interface
• Claude Code, Anthropic's command-line coding tool
• Claude Console at platform.claude.com, the developer console for the Claude platform
• Claude Cowork

There is a small discrepancy in the report itself: the incident's title names claude.ai, Claude Code and platform.claude.com, while the affected-services list adds Claude Cowork as a fourth product. Either way, the blast radius covered both the consumer chatbot and the tooling developers build on, because all of them sit behind the same account and authentication layer. A single fault at that layer was enough to produce errors across the portfolio at once.

The incident also drew attention on Hacker News, where it reached the front page, an indication of how closely technical users now track Claude's availability.

Why it matters

The outage was short, but its shape is instructive. Claude Code and the Claude platform have become everyday working tools for a large number of developers, and an authentication fault is uniquely disruptive: if you cannot sign in, every downstream capability, from chat sessions to coding agents to API work in the console, is blocked at the same time. There is no partial degradation, only absence of access.

It is also a reminder of the shared-infrastructure risk that comes with consolidating work around one AI vendor. Four distinct products failed together because they depend on one common service, and teams that have built workflows exclusively around Claude felt the outage regardless of which product they were using.

The response timing cuts both ways. Acknowledging an issue at 09:58 UTC and deploying a fix by 11:05 UTC is a fast turnaround, and the public status page gave affected users a place to check. But with the incident still in a monitoring state at the time of the last update, and no root-cause detail published, users cannot yet judge whether the fix addresses the underlying problem or its symptoms.

Until Anthropic publishes more detail, the status page remains the authoritative source on whether recovery has held, and anyone who hit sign-in errors during the window should treat it, rather than social media chatter, as the place to confirm normal service.