Rate limits
Limits keep the API fast for everybody. They are high enough for normal use, and every answer tells you how close you are.
The limits
| Limit | Per API key | All keys of your workspace together |
|---|---|---|
| Requests a minute | 1,200 | 3,600 |
Sends a minute (POST /messages and POST /email/messages) | 600 (about 10 a second) | 1,800 |
A send counts as a request and as a send. The limits use a sliding one-minute window. More keys never mean higher limits: your workspace has its own limit on top of each key's.
Headers
Every answer to a known key carries three headers. They describe the limit that is closest to running out, so following them never runs you into another one.
| Header | Meaning |
|---|---|
X-RateLimit-Limit | The most requests allowed in the window. |
X-RateLimit-Remaining | How many are left. |
X-RateLimit-Reset | When the window frees up, as Unix seconds. |
When you go over
You get 429 RATE_LIMITED with a Retry-After header (seconds to wait). details.limit says which limit you hit: requests, sends, workspace_requests or workspace_sends. A refused request uses up no slot, so waiting is enough.
429 Too Many Requests
{
"error": {
"message": "Too many requests. Wait for the Retry-After seconds.",
"code": "RATE_LIMITED",
"details": {
"limit": "requests"
}
}
}- Wait for
Retry-After, then try again. Use the sameIdempotency-Keyfor the retry. - To load many contacts, use
POST /contacts/bulk: up to 1,000 contacts in one request. - Email has its own limits too: your workspace's daily email limit (
429 EMAIL_DAILY_LIMIT) and, for transactional email, 5 a day and 20 a week for each address (429 ADDRESS_LIMIT). Both answer withRetry-After.

