Rate limits

Each API key may make up to 600 requests per minute, counted per calendar minute (UTC). Bursts are fine as long as a minute's total stays below the limit.

When a key goes over the limit, requests fail with 429 Too Many Requests until the next minute starts. The Retry-After header says how many seconds that is:

JSON
{
"type": "about:blank",
"title": "Too Many Requests",
"status": 429,
"detail": "API key rate limit exceeded. Retry later.",
"code": "RATE_LIMITED"
}

Staying within the limit

  • Use batch endpoints. records:batchUpsert and events:batch do the work of many single calls in one request.
  • Wait as long as Retry-After says. On a 429, wait that many seconds and retry. Without the header (other 429s), back off exponentially with a little jitter (for example 1 s, 2 s, 4 s…).
  • Use webhooks instead of polling. Let the CRM tell you when something changes; see Webhooks.
  • Split heavy workloads across separate keys for separate integrations. Each key has its own limit, and keys can be revoked independently.
Node.js
async function withRetry(request, attempts = 5) {
for (let i = 0; ; i++) {
const res = await request()
if (res.status !== 429 || i === attempts - 1) return res
const after = Number(res.headers.get('Retry-After'))
const wait = after > 0 ? after * 1000 : 2 ** i * 1000
await new Promise((r) => setTimeout(r, wait + Math.random() * 250))
}
}

Your organisation's limits

Separate from the request rate, each organisation has daily sending limits and a few totals: marketing and transactional email, automation emails and texts, new automation starts, SMS, file storage and the AI assistant's monthly budget. GET /api/v1/limits returns each one with what is used, the limit, when it resets and a state (ok, near from 80%, reached):

JSON
[
{ "key": "marketing_emails", "used": 4120, "limit": 5000, "period": "day",
"resetsAt": "2026-11-01T00:00:00Z", "state": "near" },
{ "key": "ai_tokens", "used": 506717, "limit": 3000000, "period": "month",
"resetsAt": "2026-12-01T00:00:00Z", "state": "ok" }
]

Nothing is lost at a daily sending limit: campaigns and automation sends carry on after the reset, and new automation starts over the limit are spread over the following days. A campaign held by the daily marketing limit waits until 08:00 in your organisation's time zone at the earliest, never in the middle of the night, and says so: waitingReason: "daily_limit" and waitingUntil on the campaign. A sending domain that is still warming up has its own daily limit (warm_up, see Email sending). A transactional email over its limit is refused (429, code send_cap_exceeded), so retry it the next day or ask Code F to raise the limit. Daily limits count the UTC day.