Michael JohnsonOn September 29, Anthropic reported an outage affecting Claude.ai, Claude apps, Claude Code, Cowork, and some API requests. The company said the impact lasted from 14:00 to 14:59 UTC. Services recovered, but some messages sent during that hour may not have been saved. At the time of checking, Claude’s status page showed all systems operational.
What happened
In addition to an elevated error rate, users may have encountered failed requests, problems loading or sending conversations, and repeated sign-in prompts. Anthropic later reported disruptions to sign-in, creating new chats, voice conversations, file uploads, and Claude Code and Cowork sessions. This is a list of affected features, not evidence that they were unavailable to every user.
According to Anthropic’s incident timeline, the company mitigated the initial errors by 14:36 UTC, but a separate issue with signing in and starting new sessions persisted. Most services recovered at 14:59. At 16:27 UTC, the incident was marked resolved, and Anthropic said services had been operating normally since recovery.
Anthropic’s warning said some messages sent between 14:00 and 14:59 UTC may not have been saved. The company did not say that all messages were lost, nor did it say whether messages that did not appear in chat history could be restored.
What users should check
If you sent messages during that time window, check your chat history. Anthropic confirmed disruptions to Claude Code, Cowork, and the API, but its published report does not specify which user data or requests may have been affected.
The company did not name the cause of the outage, the number of users affected, or the proportion of failed requests. Based on the available information, the incident cannot be attributed to a specific model, infrastructure provider, or level of demand.
The “operational” status reflects the condition of the services at the time of checking; it does not undo the reported outage. 9to5Google covered the September 29 disruption, citing Anthropic’s updates.