The limit
Each API key can make 120 requests per minute, counted in fixed one-minute windows. Keys don't share their limit, so separate systems can use separate keys.
Above the keys, one IP address can make at most 1,200 requests per minute in all, whatever keys it uses. That 429 comes from our network edge: it has no JSON body and no Retry-After, so wait a minute before you retry.
| Header | On | Value |
|---|---|---|
X-RateLimit-Limit | requests that passed the key checks | The requests allowed per minute (120) |
X-RateLimit-Remaining | requests that passed the key checks | Requests left in the current minute |
Retry-After | 429 responses | Seconds until the next window starts |
Over the limit
429 Too Many Requests
Wait for Retry-After seconds, then continue. For bulk work (importing candidates, a nightly export) spread the requests out, or watch X-RateLimit-Remaining and pause at zero.