Skip to the board
DIDRESET
UTC

"API Error: Request rejected (429) · this may be a temporary capacity issue. If it persists, check https://status.claude.com."

by vortwangUpdated

"API Error: Request rejected (429) · this may be a temporary capacity issue. If it persists, check https://status.claude.com." means "You have hit the rate limit configured for your API key, Amazon Bedrock project, or Google Cloud project", per the Error reference. It is not the five-hour session or weekly limit of a claude.ai subscription, so no reset time is shown and there is no account-wide reset to wait for; the board row below is for subscribers.

Last reset
04 SEP 20:08Z

@lydiahallie

Status
OVERDUE

The gap since the last reset is past 1.5x the recent median.

Next est.
ANY TIME

Days since
14.4

days

What does "API Error: Request rejected (429) · this may be a temporary capacity issue. If it persists, check https://status.claude.com." mean?

The Error reference files this message under a different cause from the subscription limits: the rate limit of your API key, Amazon Bedrock project or Google Cloud project. The trailing sentence "names where to check service health and varies by provider": Bedrock, Google Cloud's Agent Platform and Microsoft Foundry name their own status pages, and a custom ANTHROPIC_BASE_URL names the gateway host.

Two neighbours look similar. "API Error: Server is temporarily limiting requests (not your usage limit)" is a short-lived server-side throttle that Claude Code retries with backoff. "You've hit your session limit · resets {time}" is the subscription window, which carries a reset time. The 429 message carries neither a reset time nor a link to /usage-credits, which is how you can tell it apart.

A common trap is being on a Pro or Max plan and still seeing this: "A stray ANTHROPIC_API_KEY in your environment can route requests through a low-tier key instead of your subscription." /status shows which credential is active.

Which window is full?

Which window is full?
WindowResetsWhere to check
API key / cloud project rate limitPer-minute and per-day tiers set in your provider console; no session or weekly reset clock/status shows the active credential; the Claude Console or your cloud console shows the limits
Subscription session and weekly limitsNot involved: this message does not come from a claude.ai subscription window/usage (only meaningful when a claude.ai login is the active credential)

What to do when you see it

  1. Run /status

    Confirm the active credential is the one you expect. If you meant to use a Pro or Max subscription, unset ANTHROPIC_API_KEY in the shell, remove it from your shell profile, relaunch claude, and /login if needed.

  2. Check the status page

    The message points to status.claude.com on the Anthropic API; Bedrock, Google Cloud and Foundry configurations name their own status page instead.

  3. Check your provider console

    For Anthropic API keys, the rate limits reference explains how tiers work and how to set per-workspace caps; request a higher tier if you hit it routinely.

  4. Reduce concurrency

    Lower CLAUDE_CODE_MAX_TOOL_USE_CONCURRENCY, avoid running many parallel subagents, or switch to a smaller model with /model for high-volume scripted runs.

When does it reset?

There is no session or weekly reset to wait for. API rate limits are per-minute and per-day tiers on your key or cloud project, so the message carries no reset time; when the limit is a temporary capacity issue, retrying after a short pause is what the docs suggest.

Account-wide resets on this board apply to claude.ai subscription windows. If /status shows an API key as the active credential, those resets do not affect you; switch to your subscription login first if that is what you pay for.

Related errors

FAQ

I am on Max. Why do I get a 429 with no reset time?

Most likely a stray ANTHROPIC_API_KEY is routing requests through an API key instead of your subscription. The docs say to run /status and check the API key row, unset the variable, remove it from your shell profile, relaunch claude and /login.

Is this the same as "Server is temporarily limiting requests"?

No. That message is a short-lived server-side throttle unrelated to any quota and is retried automatically. "Request rejected (429)" is the configured rate limit of your API key or cloud project; the fix is a higher tier, less concurrency, or the right credential.

Does a global reset on the board help?

No. Board resets refresh claude.ai subscription windows. API key and cloud-provider usage is metered per request against your tier, and the Interactive mode docs say there is no reset to wait for on those credentials.

Why does the message name a different status page?

The trailing sentence varies by provider: Amazon Bedrock, Google Cloud's Agent Platform and Microsoft Foundry configurations name that provider's service status instead of status.claude.com, and a custom ANTHROPIC_BASE_URL names the gateway host.

Sources

  1. Claude Code docs: Error reference — "Request rejected (429)"verified 2026-09-18
  2. Claude Platform docs: API errors (401 authentication_error, 429 rate_limit_error)verified 2026-09-18
  3. Claude Platform docs: Rate limits (tiers, per-workspace caps)verified 2026-09-18
  4. Claude Code docs: Error reference — "Server is temporarily limiting requests"verified 2026-09-18
  5. Claude Help Center: Use Claude Code with your Pro or Max plan (/status, /login, shared limits)verified 2026-09-17
  6. Anthropic status page (status.claude.com)verified 2026-09-18

Want the next Claude reset on your phone?

Get alertedPage me

Sponsors