Claude outage on September 29 hit apps and developer services
Published by AINave Editorial • Reviewed by Ramit
Anthropic’s September 29 partial outage affected Claude, Claude Code, Claude Cowork, the Claude API and Console. The initial errors were mitigated within 36 minutes, but that did not mean every workflow was back: a separate issue continued to disrupt sign-ins and other actions for many users, according to the incident update reported by 9to5Google.
The first incident and the follow-on problem were different
The reported errors began at 14:00 UTC and were mitigated at 14:36 UTC. Existing conversations were mostly working again, but the follow-on problem still prevented many users from signing in through single sign-on or Sign in with Apple, starting new chats, using voice conversations, opening Claude Code or Cowork sessions, making purchases, and uploading files. Anthropic had said it was aware of the outage shortly before 10:30 a.m. ET; the article does not give a final resolution time for the remaining issue. These status details describe the incident as reported, not Claude’s availability now.
That distinction matters for teams using Claude in more than one way. A mostly working existing conversation did not guarantee that a new session, file upload or developer workflow would work. The report identifies the affected services and actions, but gives no technical root cause, so it does not establish whether the problems shared one underlying failure.
A practical status note for users
Anthropic’s status message advised people who were already signed in not to sign out while a fix was underway. That warning fits the shape of the disruption: users could retain access to existing conversations while facing trouble starting new work or authenticating again. The article does not report when that advice stopped applying. The recommendation was specifically for users already signed in.
The article also noted earlier partial Claude outages on September 22, 15 and 11, and a September 3 outage that coincided with incidents affecting ChatGPT and Grok. Those dates put the latest incident in context, but they do not explain its cause. DownDetector reports rose during the September 29 event; the article said limited reports of ChatGPT and Gemini trouble did not appear accurate, and neither OpenAI nor Google had confirmed an issue. The report does not establish a shared outage across those services.
For operators, the clearest lesson from this incident is about status granularity: mitigation of an initial error is not the same as full recovery across authentication, new sessions, uploads and developer tools. The September 29 update documented that gap, but not how long it lasted.
