Anthropic service status
Check whether the provider lists a problem with the feature you use. Look at the update time and any affected regions.
Live Domain Check
Run a live reachability check for anthropic.com. Compare the regional responses with the problem you see in Anthropic.
Find Anthropic help and reporting links, specific checks, and step-by-step guidance below the live result.
Checking reachability from multiple regions.
Share a quick update and see what other visitors are experiencing.
What’s happening? Choose one to send your report.
Loading community reports…
—problem reports
—working reports
No problems reported in this period.
Visitor reports are separate from our automated checks and don’t confirm an outage. How reports work · Privacy
The anthropic.com company website is separate from Claude chat and authenticated API traffic.
The checker above requests anthropic.com from eight configured locations and shows the responses and measurement times. Compare those observations with the actual Anthropic action that failed. A homepage response alone cannot verify an account, app workflow or transaction.
Distinguish the Claude application from the Anthropic website when recording the failed action: Claude status check. Compare results separately; these links do not establish a shared outage.
Looking for another service? Browse AI service checks.
Check whether the provider lists a problem with the feature you use. Look at the update time and any affected regions.
Choose Claude chat or developer-platform support according to the failing product.
Open the website in your browser and compare it with the result above.
Open Anthropic help / reporting. Use the provider's customer-service or account-help route for the affected product. Sign-in may be required for private account questions.
Affected product, endpoint/model, request ID, error type and UTC time; never include an API key or confidential prompt.
Attach the checker’s downloaded JSON if it helps show the hostname, regional responses and original timestamps. Describe what happened in your browser or app as well. A report sent to WebsiteDown.org is not automatically forwarded to Anthropic.
These are troubleshooting comparisons, not claims that Anthropic is currently experiencing a particular incident.
Identify whether the issue concerns the company site, claude.ai or an API endpoint, then use the corresponding status component.
For an API failure, capture the response status and request identifier; compare rate-limit errors with provider incident information.
A 403 Forbidden or 429 Too Many Requests means the host answered but refused or limited that request. A timeout gives no HTTP response. Keep those outcomes separate and check the available-region count before treating the result as broad evidence.
Use these local troubleshooting steps after the down-check workflow when Anthropic seems broken only for you. This section focuses on app, browser, account, and network fixes.
Open anthropic.com in your current browser, then test in a private window or second browser. If only one session fails, compare its account, cookie and extension context before assuming the entire service is unavailable.
Keep an existing working session and preserve unsaved work. For a sign-in failure, use the official account-help route. Avoid repeated password or security resets until you confirm this is not a broader Anthropic issue.
If possible, compare Wi-Fi with mobile data and note which one fails. Leave any required work VPN or security settings in place.
Save timestamp, device, network type, exact error, final URL, and status code. Use the check workflow above before contacting Anthropic support.
Run the anthropic.com check and compare its original timestamps and responses with your actual error. The anthropic.com company website is separate from Claude chat and authenticated API traffic.
Use the community buttons on this page to share whether the service works for you. For help from Anthropic, use the official help/reporting resource linked above. Choose Claude chat or developer-platform support according to the failing product.
Identify whether the issue concerns the company site, claude.ai or an API endpoint, then use the corresponding status component. For an API failure, capture the response status and request identifier; compare rate-limit errors with provider incident information.
Affected product, endpoint/model, request ID, error type and UTC time; never include an API key or confidential prompt.