Is Claude API down

Is Claude API Down? Anthropic Status and Error Checks

Honest boundary: llmping does not ping user API keys and does not claim fake real-time provider status. Use this page to separate outage symptoms from key, quota, permission, and runtime failures.

What this page can check

Check whether a Claude or Anthropic API failure looks like provider overload, network reset, invalid auth, or an application-side retry problem.

Official provider status page: https://status.anthropic.com/

Historical public endpoint uptime

llmping samples provider public endpoints and official status feeds every 5 minutes through a Cloudflare Worker. These percentages never use user API keys and should be read as public reachability evidence, not a guarantee for every authenticated workload. The latest check time appears with the data; if it is more than 15 minutes old, treat the panel as historical rather than current.

Loading real uptime samples...

Immediate checks

  1. 529 overloaded_error points to provider capacity pressure.
  2. ECONNRESET is a transport failure and can come from proxies or serverless timeouts.
  3. 401/403 class errors are normally account, key, or permission issues.
  4. Compare one short request against one streaming request before blaming the provider.

cURL test command

curl -i https://api.anthropic.com/v1/messages \
  -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{"model":"claude-3-5-haiku-latest","max_tokens":16,"messages":[{"role":"user","content":"ping"}]}'

Run this from your own terminal or server. A status code is only useful after you know whether auth, quota, or endpoint selection is correct.

Run diagnosis

Related errors

Markdown version

Optional site analytics

Allow anonymous usage counts with Umami. Error text, API keys, prompts and generated commands are never sent. The tools work with analytics off. Analytics settings

Optional Umami analytics is off.