Standard Compute

All fixes · OpenClaw · 429 · Rate limits

OpenClaw “API rate limit reached” (429)

API rate limit reached

OpenClaw shows “rate limit reached” when its model provider returns a 429. The same message can come from several sources (provider RPM/TPM limits, a provider in cooldown, or no fallback configured), but it almost always traces back to an upstream provider throttling you.

What causes it

  • The upstream model provider (OpenAI, Anthropic, OpenRouter) hit its per-minute or quota limit.
  • Background load: OpenClaw heartbeats, retries, and parallel tool calls consume the provider budget even when you’re idle.
  • A provider is in cooldown and no fallback is configured, so requests keep failing.
  • Oversized agent sessions that fan out many provider calls per step.

How to fix it

  1. Configure fallback models in openclaw.json so OpenClaw reroutes when one provider is throttled.
  2. Reduce heartbeat frequency and concurrency to lower background load.
  3. Check provider status with your model dashboard, and upgrade the provider tier for higher limits.
  4. Use a provider with higher per-minute limits so background activity can’t exhaust them.

Running an agent? OpenClaw’s heartbeats and retries are the usual culprit. They can burn a large share of a per-minute quota before you send a single message.

Compare APIs for OpenClaw

Best API for OpenClaw →
5 options compared, with prices and limits.

Keep working on a fixed monthly plan

Point OpenClaw at Standard Compute (base URL https://api.stdcmpt.com/v1, model “standardcompute”): each key has a high request ceiling (300 requests a minute on Starter, 2,000 on Pro), and transient upstream errors are retried with fallback to another provider for the same model, so you don’t hand-configure fallbacks.

OpenClaw custom provider
baseUrl  https://api.stdcmpt.com/v1
api      openai-completions
apiKey   YOUR_API_KEY
model    standardcompute
Get your API key OpenClaw setup →

Plans from $19 a month with $19 of compute. Eligible new accounts can try $0.25 of compute first, no card. Trial terms

FAQ

Why does OpenClaw hit rate limits when I’m not even using it?

Background heartbeats, transient-failure retries, and parallel tool calls keep making provider requests. On a small per-minute quota, that background load alone can trigger 429s.

What’s the most reliable fix?

Configure fallbacks so a throttled provider doesn’t stall everything, or use a provider with higher per-minute limits and built-in failover, so fewer 429s reach OpenClaw.

Related errors