Rate limits
Current limits (V0)
Requests are rate limited per API key per minute. Limiting is protective rather than an exact quota: it exists to keep any one key from overwhelming the API, thresholds are approximate and set generously above normal integration traffic, and they are tuned as real usage grows — a well-behaved integration should never see them.
When a request is throttled, it receives:
{
"error": {
"type": "rate_limit_error",
"code": "rate_limit_exceeded",
"message": "Rate limit exceeded"
}
}
with HTTP status 429 and a Retry-After header stating how many seconds to wait before
retrying. Application-level budgets are independent per key — throttling one key does not
consume another key's budget. Edge IP limits, by contrast, are shared by everything behind
one IP address (see below).
Separately, infrastructure-level IP rate limiting applies at the network edge, per client
IP address across all requests. When exceeded, the edge layer may block further requests from
that IP, and those blocks may not use the standard Nory { "error": ... } envelope.
Recommended client behaviour
- Honour
Retry-After— on a429, wait the indicated seconds before retrying rather than retrying immediately. - Use pagination — fetch data in pages rather than tight polling loops.
- Cache where appropriate — reference data such as locations changes infrequently.
- Backoff on errors — use exponential backoff on
5xxor connection errors. - Shared NAT — multiple services behind one IP share one edge rate-limit quota.
Need higher limits?
Contact your Nory account manager if your integration requires sustained high request volumes.