Rate limits & quotas

Request rate limit

The public API allows 10 requests per second per workspace by default. All API keys of a workspace share the same budget, and it covers every /api/v1 endpoint, MCP tools/call requests and SMTP messages. Every /api/v1 response carries the current state:

ratelimit-limit: 10
ratelimit-remaining: 7
ratelimit-reset: 1

ratelimit-reset is the number of seconds until the window resets. Over the limit, the API answers 429 rate_limit_exceeded with a retry-after header (seconds). Need more throughput? Use batch sends (up to 100 emails per request) or contact support.

Email quotas

Quotas count recipients, not API calls: an email to 3 addresses in to plus 2 in cc uses 5. Every accepted transactional email counts from the moment it is accepted (queued, scheduled, sent, delivered, bounced, …); only emails that end failed or canceledgive their quota back. The check and the insert are atomic, so parallel requests can't overshoot.

Plan limits are listed on the pricing page. Broadcast emails don't count toward the transactional quota — Marketing plans limit contacts instead (see Broadcasts).

Pay-as-you-go

On a paid plan with pay-as-you-go enabled, sending continues past the monthly quota and the overage is billed in whole buckets of 1,000 emails, rounded up. Two guards stop runaway costs: an overage cap (by default five times the plan quota) and an optional spend limit you set in Settings → Usage. With pay-as-you-go disabled, sending stops at the quota.

Scheduled and batch sends

• A scheduled email counts toward the period it is scheduled to be sent in, and is checked again when it goes out.
• A batch is all-or-nothing: if the batch's recipients don't fit, no email in it is accepted.

Other limits