Is Claude down right now? Check it live, then keep working.
The status here comes straight from Anthropic and refreshes every minute. APIsRouter reaches Claude over more than one route, and 99.2% of requests through it succeed.
Claude outages in the last 30 days
Each incident Anthropic posted in the last 30 days, with how long it lasted and what it touched, loads here from status.claude.com.
Keep working while Claude is down
Many Claude incidents hit one model or one path, not everything at once. Three ways around them:
- Route around a failing path: APIsRouter reaches Claude over more than one route. A request that times out or hits a gateway error on one route is retried on another before an error reaches you; 99.2% of requests through APIsRouter succeed.
- Switch Claude models in Claude Code: When an incident names one model, run /model in Claude Code and pick another Claude model. Your key and settings stay the same.
- Switch to GPT or Gemini in other apps: Apps that call the OpenAI API can change the model ID to gpt-6-sol or gemini-3.8-flash on the same APIsRouter key, and switch back when Claude recovers.
What Claude API errors mean
The code tells you whether to wait, retry or change something on your side.
| Error | What it means | What to do |
|---|---|---|
| 529 overloaded_error | Anthropic's API is overloaded for everyone, not just you. | Retry after a short wait, or switch to another model until it clears. |
| 500 api_error | Something failed inside Anthropic's systems. | Retry with exponential backoff. If it keeps failing, check the live status. |
| 504 timeout_error | The request ran too long and timed out. | Stream long responses instead of waiting for one reply. |
| 429 rate_limit_error | Your organization hit a rate limit or its monthly spend cap. | Wait for the retry-after time, or spread requests out. A spend cap keeps failing until it resets. |
Run Claude Code through APIsRouter
- Get a key: Start free on a $2 trial credit, or top up from $10 below. A top-up key arrives by email.
- Point Claude Code at APIsRouter: Add these lines to ~/.claude/settings.json.
- Run claude: Signed in with Pro or Max? Run /logout first so Claude Code uses your key.
{
"env": {
"ANTHROPIC_BASE_URL": "https://api.apisrouter.com",
"ANTHROPIC_AUTH_TOKEN": "sk-your-apisrouter-key",
"ANTHROPIC_MODEL": "claude-opus-5-5",
"ANTHROPIC_DEFAULT_OPUS_MODEL": "claude-opus-5-5",
"ANTHROPIC_DEFAULT_SONNET_MODEL": "claude-sonnet-5-5",
"ANTHROPIC_DEFAULT_HAIKU_MODEL": "claude-haiku-4-5-20251001"
}
}Two ways to pay
Both fill one balance, spent per token on Claude, GPT, Gemini and the rest.
FAQ
Is Claude down right now?
The live status at the top of this page comes from Anthropic's status page, status.claude.com, and refreshes every minute. If it shows Claude working but your requests still fail, check the error code: some errors come from your plan or your requests, not an outage.
Why does Claude say it is overloaded?
An overloaded message or error 529 means Anthropic's API is under heavy traffic across all users. It is not your account. Wait a moment and retry, or switch to another model until it clears.
Is Claude Code down?
Claude Code has its own line in the live status. It calls the same Claude API, so an API incident usually reaches Claude Code too. On an APIsRouter key, a request that times out or hits a gateway error is retried on another route first.
Claude says I have reached my usage limit. Is it down?
No. That is your plan's limit: Pro and Max usage resets every 5 hours, with a weekly cap. Claude Code on an API key has no usage windows; you pay per token instead.
How do I keep using Claude when it is down?
Let APIsRouter retry timeouts and gateway errors on another route, switch to another Claude model with /model in Claude Code, or move an OpenAI-compatible app to GPT or Gemini on the same key.
Does APIsRouter go down when Anthropic does?
APIsRouter reaches Claude over more than one route and retries timeouts and gateway errors on another, so a failure on one route does not have to reach you. When Anthropic is overloaded (error 529) or a model fails everywhere, requests to it fail here too; switch models until it recovers.
Where can I see past Claude outages?
The list on this page shows the incidents Anthropic posted in the last 30 days, how long each lasted and what it touched. Older incidents are on status.claude.com/history.