---
title: Sending limits
description: The protections that can refuse or hold your email, what triggers them, what you see, and how to recover.
---

BillionSend protects your sender reputation and the shared sending infrastructure with a few automatic limits. Most of them never trigger for a healthy account. This page lists each one: what triggers it, what you see, and how it ends.

An email that hits a limit has one of two outcomes:

- **Refused.** The request fails right away with an error and nothing is stored. The [SMTP relay](/docs/smtp) refuses the same mail with the matching SMTP error.
- **Held.** The email is accepted and shows as **Waiting** in the dashboard, with the reason on the email's page. In the API it stays `queued`. It goes out by itself once the limit lifts.

## Your plan's volume

| Plan | Limit | Past the limit |
| --- | --- | --- |
| Free | 100 emails per UTC day | Up to 150 still go out. Beyond that, mail is held until the next UTC day. When 300 are waiting, new sends get `429 daily_quota_exceeded`. |
| Starter, Pro and Scale | Included monthly volume | With **overage** on (the default), extra emails are billed at $0.40 per 1,000, and sending stops at 5 times the included volume. With overage off, sending stops at the included volume. Either way the API answers `429 monthly_quota_exceeded` until the period renews. |

- Overage pauses while a renewal payment is past due. A paid plan whose period ended 7 days ago without renewing is limited like Free.
- Owners get an email, and your [webhooks](/docs/concepts/webhooks) get `quota.warning`, `quota.reached` and `quota.paused`, as the team nears and reaches its limit.
- To lift it: upgrade, turn overage on in **Settings → Billing**, or wait for the period or day to reset.

See [Plans and limits](/docs/billing) for the full price and limit tables.

## Bounce and complaint rates

If too many of your emails bounce or are marked as spam, mailbox providers distrust your mail, and BillionSend steps in before they do. Only hard bounces count, and complaints count only for providers that report them. Gmail, Apple and Deutsche Telekom do not.

| | Warning | Pause |
| --- | --- | --- |
| Hard bounce rate | 4% or more over the last 7 days | 5% or more over today and yesterday |
| Complaint rate | 0.05% or more over the last 7 days | 0.1% or more over today and yesterday |
| Needs at least | 100 sends, and 1 bounce or 2 complaints | 100 sends, and 10 bounces or 3 complaints |

- **Warning.** Nothing is refused. Broadcasts slow down to one email per second, a banner shows in the dashboard, and plan upgrades are unavailable until the rates drop.
- **Pause.** All new sends are refused with `403 sending_paused`, from the API and the SMTP relay. Emails already waiting are held.
- **How it ends.** Automatically, as soon as the rates fall back under the lines. There is nothing to request. Clean your list, stop mailing addresses that bounce, and send only to people who asked for your mail.
- **How you hear about it.** Owners get an email and a webhook, `deliverability.warning` or `deliverability.paused`. The **Deliverability** page shows your rates, and the [account score](/docs/concepts/deliverability-insights) never reads higher than the warning or pause allows.

## Platform pause on broadcasts

Amazon SES judges the whole platform by the combined rates of all senders, and one bad list can affect everyone. When those rates near the line SES acts on, new broadcasts are refused with `403 broadcasts_paused`, and scheduled broadcasts wait. Transactional email keeps flowing. It clears on its own once the rates recover.

## New domain warm-up

A domain registered less than 30 days ago sends at a daily cap that grows with its age: 100, then 300, then 2,000 emails a day. Mail over the cap is held and goes out the next day, and when too much is waiting the API answers `429 daily_quota_exceeded`. Upgrading does not lift it, and a domain that sends cleanly moves up sooner. See [Domains](/docs/concepts/domains) for the table.

## Recipient checks

Before sending, BillionSend leaves out recipients that cannot or should not get the mail:

- **Suppressed addresses.** Addresses that hard bounced, complained, were added by you, or were found invalid by verification. See [Suppressions](/docs/concepts/suppressions).
- **Unsubscribed contacts.** Broadcasts and topic mail skip people who unsubscribed. See [Unsubscribes](/docs/concepts/unsubscribes).
- **Addresses that can never receive mail.** Reserved or example domains, domains that do not exist, domains with no mail server, and disposable-email domains are refused with `422`. These checks are free and do not count toward your volume. They also keep such addresses out of contact imports. To test sending without emailing anyone, use `success@simulator.amazonses.com`.

When every recipient is left out, the send answers `422 all_recipients_suppressed`.

## Request limits

| Limit | Value |
| --- | --- |
| API requests per API key | 600 per minute |
| API requests per team, all keys together | 3,000 per minute |
| Recipients per email (`to`, `cc` and `bcc` together) | 50 |
| Attachments per email | 10 MiB in total |
| Emails per `POST /emails/batch` | 100 |

Past a rate limit, the API answers `429 rate_limit_exceeded` with a `retry-after` header. Wait that many seconds and retry.

## Safety checks and account holds

- **Held for a safety check.** Some emails are checked automatically before they go out, most often from a new account. The email page shows **Held for a safety check**. It is released or cancelled after the check. If a hold lasts 7 days, the email fails with `email.failed`.
- **Account under review.** Your account may be put under review. A banner shows, new broadcasts are refused, and mail is accepted but held. Contact support.
- **Account suspended.** All sends are refused with `403 team_suspended`, and a red banner shows. Contact support.
- **Amazon SES pause.** If SES pauses your team because its rates climbed, owners get an email with SES's reason. Your mail is held, and goes out by itself when SES enables sending again.

## See why an email is held

Open the email in the dashboard. The page names the reason: warming up the domain, plan limit, sending paused, platform capacity, a safety check, and so on. The API only shows `queued`, so check the dashboard or the [Deliverability](/docs/concepts/deliverability-insights) page when mail is slow to leave.
