Claude.ai, the Claude API and console.anthropic.com broke for two hours and eight minutes on 29 September 2026. The fault sat on Anthropic's side, affecting sign-ins, new chats, Claude Code and Cowork sessions, and some API requests.
Started
14:21 UTC
Duration
Lasted 2h 8m
Source
IsDown
Next time
Anthropic is degraded. Is your product working?
An incident on their side does not always mean an outage on yours, and the only way to know is to already be watching your own endpoints. Set that up in under a minute and you will be notified the next time your own site is affected. And there will be a next time.
50 monitors for your own websites. No card.
Affected components: claude.ai, console.anthropic.com, api.anthropic.com
What happened?
Anthropic's own status page logged one incident starting at 14:21 UTC, covering claude.ai, api.anthropic.com and console.anthropic.com, and it ran for 2h 8m before resolution at 16:27 UTC. The first update described elevated error rates on Claude.ai, the desktop and mobile apps, Claude Code and Claude Cowork, with failed requests, errors loading or sending conversations, and users being asked to sign in again. Seven minutes later Anthropic added that some requests to the Claude API were also returning errors. A mitigation applied around 14:41 UTC cut error rates substantially, but sign-in, account actions and Claude Code or Cowork sessions kept failing. At 15:00 UTC Anthropic described a second, distinct issue: single sign-on and Sign in with Apple were unavailable, and new chats, voice conversations, purchases and file uploads were blocked, while telling users who were signed in not to sign out. Most services recovered by 14:59 UTC according to the 15:11 UTC update, though Anthropic flagged that some messages sent between 14:00 and 14:59 UTC may not have been saved. The resolved update at 16:27 UTC stated that impact ran from 07:00 to 07:59 PT, 14:00 to 14:59 UTC, which is a shorter window than the full incident record. One user report from Ukraine at 14:27 UTC said simply "This response didn't load error," matching the errors Anthropic itself described. No post-mortem explaining the underlying cause has been published.
Learning
UptimeRobot's 30 day record for Anthropic shows 13 incidents, all resolved, with a median duration of 1h 29m, so an outage of this length is not unusual for this provider. If your own uptime checks against Anthropic's endpoints stayed green through this window, that is not a contradiction: a check can see a reachable endpoint while real requests still fail on sign-in, session handling or save operations further inside the stack, which is exactly what Anthropic's own updates describe here. This incident involved two separable failures layered together, generic elevated errors first, then a sign-in and session problem that outlasted the first fix, and a monitor pinging a single URL would not distinguish between them. Anthropic's 90 day history logs 86 incidents with a median duration of 1 hour 7 minutes, so this is a provider with a frequent pattern of short, recoverable breaks rather than one clean event. Watching your own endpoints tells you whether your product is up, it does not tell you whether Anthropic is the reason it isn't. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.
