Rate Limits

Inflow limits how quickly you can call the API with a secret API key. Rate limits protect platform stability and keep response times predictable for every integration.

Default Limits

Both windows apply on every request. Hitting either limit returns a 429 response.

If your integration needs higher throughput, contact Support — limits may be increased on request if needed.

Public and SDK Keys

This rate-limit policy applies to requests authenticated with a secret API key. Public / SDK keys are not subject to the same limits. Other protections may still apply at the infrastructure layer.

When You Hit the Limit

Exceeded limits return HTTP 429 Too Many Requests. The response body looks like:

{
  "status": "error",
  "message": "Rate limit exceeded: max 7 requests per second"
}

When the per-minute window is exceeded, the message refers to the minute limit instead (max 120 requests per minute).

Rate-limit responses for secret API keys do not currently include X-RateLimit-* or Retry-After headers. Use the message field and HTTP status to decide how to back off.

How to Handle 429 Responses

1. Detect 429 before treating the call as a permanent failure.
2. Back off before retrying:
   • Per-second limit → wait about 1 second, then retry.
   • Per-minute limit → wait about 60 seconds, then retry.
3. Prefer exponential backoff with jitter if you retry in a loop.
4. Avoid bursts and tight polling loops; space out requests under the limits above.

Do not run load tests against the sandbox. Sustained high volume can exhaust your quota and may lead to temporary API access suspension.

Best Practices

• Stay under both windows — design clients for ≤ 7 req/s and ≤ 120 req/min under normal load.
• Retry 429s with backoff — do not hammer the API immediately after a rate-limit response.
• Prefer webhooks over polling — reduce unnecessary GET traffic; see Setting Up Webhooks.
• Monitor 429 rates — a rising 429 count usually means you need pacing or a limit increase.
• Ask if you need more — contact [email protected] if your traffic pattern requires higher limits.