Official provider signal

Is Claude down?

Claude apps, developer platform and API from Anthropic.

Operational

Current status

Components

Official timeline

incidentElevated errors for multiple models
monitoring
Timeline →
incidentIssues with Google Play subscriptions
identified
Timeline →
incidentIntermittent error spikes for Claude Mythos 5.1 and Claude Fable 5.1
monitoring
Timeline →
incidentElevated errors for Claude Mythos 5.1 and Claude Fable 5.1
monitoring
Timeline →
incidentDegraded functionality for Claude Cowork on Windows
identified
Timeline →
incidentElevated latency on the Claude API for some users
identified
Timeline →
incidentElevated errors for multiple models
monitoring
Timeline →
incidentElevated errors for Claude Sonnet 5
monitoring
Timeline →
incidentElevated errors for Claude Sonnet 5
investigating
Timeline →
incidentDelays in credit purchases
monitoring
Timeline →
incidentDegraded performance on platform.claude.com and Claude for Microsoft Office 365
monitoring
Timeline →
incidentDegraded performance on Claude for Microsoft Office 365
investigating
Timeline →
incidentDegraded performance on claude.ai and Claude Code
investigating
Timeline →
incidentDegraded performance on claude.ai
monitoring
Timeline →

Data confidence

Operational means no issue is currently reported by the provider. It is not proof that every user or region is unaffected.

While this is open

Read the incident timeline. The provider updates say what is affected and what they have already ruled out.
Match the affected components against the calls your system actually makes before assuming this incident is yours.
Hold off on rolling back your own deploy. A rollback during a provider incident removes a variable you will want later.
Queue or fail over the work if you can. Aggressive retries against a degraded provider usually make the backlog worse.
Watch the timeline rather than polling the API. It updates as the provider posts, including the resolution.