Limits, sandbox and compliance
Credit levels, the sandbox, hourly and monthly limits, overage and the spending cap, automatic pauses, content scanning, acceptable use and data retention.
To keep delivery good for everyone, every stream has sending limits, free-plan streams start in a sandbox, and is paused automatically when its bounce, complaint or spam levels get too high. This page explains those rules and the sending policy.
Credit levels
Every account has a credit level from Lv0 to Lv6. It applies to the whole account, across all its streams.
| Level | Free plan: emails per month | Free plan: hourly limit | Paid plans: hourly limit factor | Requirements (all must be met) |
|---|---|---|---|---|
| Lv0 | 0 | 0 | Can't send | Set after a virus was sent, or by an administrator as a penalty |
| Lv1 | 100 | 20 | ×0.5 | Default for new accounts |
| Lv2 | 300 | 50 | ×0.75 | Verified account email, at least 1 verified sending domain, account at least 7 days old |
| Lv3 | 600 | 100 | ×1 | Account at least 30 days old, at least 500 emails delivered in total, last 30 days: hard bounces under 3% and complaints under 0.05% |
| Lv4 | 1,000 | 200 | ×1.25 | Account at least 60 days old, at least 2,000 emails delivered in total, at least one paid bill |
| Lv5 | 2,000 | 300 | ×1.5 | Account at least 90 days old, at least 10,000 emails delivered in total, last 60 days: hard bounces under 2% and complaints under 0.03% |
| Lv6 | 3,000 | 500 | ×2 | Account at least 180 days old, at least 50,000 emails delivered in total, last 90 days: hard bounces under 1.5% and complaints under 0.02% |
What the level does:
Free plan: the level sets the monthly volume and the hourly limit. An account can have only one free stream. When the month's allowance is used up, sending stops until next month.
Paid plans: you get the plan's included volume, not the level allowance. The hourly limit is the plan's hourly limit multiplied by the level factor. For example, Basic (2,000 an hour) allows 1,000 an hour at Lv1 and 4,000 at Lv6.
Marketing streams need Lv2 or higher.
Lv5 and Lv6 get 20% more tolerant automatic-pause thresholds (see below).
Promotion is checked daily. You move up at most one level per day.
Demotion: if any of your streams is paused automatically (see below), your account drops one level, but never below Lv1. Sending a virus drops it to Lv0. After a demotion, your level doesn't rise automatically for 30 days.
The Trust level card on the Email sending page shows your current level, how much of this month's free allowance you've used (free plan), and what you still need for the next level.
Sandbox
Streams on the free plan start in the sandbox, shown by the Sandbox tag. Streams on a paid plan skip it, and upgrading a free stream to a paid plan lifts it at once. In the sandbox:
- The stream can send at most 100 emails a day.
- It can only send to addresses at your own verified sending domains, and to addresses you've verified on the Sandbox recipients tab.
To verify a test recipient, open the Sandbox recipients tab, enter the address and click Add. A confirmation email is sent to that address. The address can receive sandbox mail once its owner clicks the link. You can verify up to 50 addresses.
On the free plan, the sandbox lifts automatically for all your streams when your account reaches Lv2. If the stream shows a Leave sandbox button, you can also lift it right away yourself. If you need the sandbox lifted for another reason, contact support.
Sending limits
| Limit | Value | When it's reached |
|---|---|---|
| Hourly limit | Paid plans: Basic 2,000, Pro 10,000, Enterprise 50,000 per hour, multiplied by your level factor. Free plan: set by your credit level. | Counted per clock hour. Further mail is refused with 452 4.7.1 (SMTP) or rate_limited_hourly (API) until the next hour. |
| Monthly volume | Paid plans: the plan's included emails. Free plan: your level allowance. | Paid plans continue with overage billing. On the free plan, sending stops until next month. |
| API request rate | 600 requests per minute per API key | 429 rate_limited |
| Recipients per API call | 50 (to + cc + bcc) | 422 invalid_request |
| Messages per batch call | 500 | 400 invalid_request |
| API attachments | 10 files, 10 MB in total | 422 invalid_request |
| SMTP message | 25 MB, 1,000 recipients per transaction | 452 4.5.3 for too many recipients |
| API keys | 20 active keys per stream | Revoke unused keys |
Each recipient copy counts as one email. Rejected recipients don't count.
Overage and the spending cap
On paid plans, emails beyond the monthly volume are billed as overage at the end of the usage period, per started block of 1,000 emails: Basic ¥0.80, Pro ¥0.60, Enterprise ¥0.40.
To protect you from runaway costs, each plan has a cap on unpaid overage: ¥200 (Basic), ¥1,000 (Pro) and ¥5,000 (Enterprise). When the cap is reached, an overage bill is issued right away and sending stops (550 5.7.1 monthly sending quota exhausted). Sending resumes once you pay the bill. Sending also stops if an overage bill stays unpaid for more than 3 days.
The Overview tab shows this month's usage and what's left of the current hour.
Automatic pauses
Each stream is checked over a rolling 24-hour window:
| Trigger | Result |
|---|---|
| Hard bounce rate above 5% (6% at Lv5 and Lv6), with at least 200 emails sent in 24 hours | Stream paused |
| Complaint rate above 0.1% (0.12% at Lv5 and Lv6), with at least 200 emails sent in 24 hours | Stream paused |
| 5 or more messages rejected as spam within 24 hours | Stream paused |
You get a warning notification when a stream is getting close to the bounce or complaint threshold. When a stream is paused, it shows Paused with the reason, you're notified, SMTP answers 451 4.7.1 Stream is paused, and the API returns stream_paused. Your account's credit level also drops by one.
To resume a paused stream, fix the cause (remove invalid addresses, check where your recipient list came from, stop the content that triggered spam rejections) and then contact support.
Content scanning and viruses
Every message, sent over SMTP or the API, is scanned for viruses and spam before it's queued.
- Spam: a message that scores as spam is refused with
554 5.7.1and counts as a strike toward the automatic pause above. - Virus: a message that contains a virus is refused, and the stream is banned immediately. All of its API keys are revoked, you're notified, and your account drops to credit level Lv0. Only support can lift a ban. After it's lifted, create new API keys.
- Marketing streams only accept mail after a successful scan. If the scanner is temporarily unavailable, marketing mail is refused with
451 4.3.0so your mail server or application can retry. Transactional mail is still accepted.
Acceptable use
By using Transactional & Marketing Mail you agree to send only mail your recipients expect:
- Consent: send marketing mail only to people who opted in to receive it from you. Don't use purchased, rented, scraped or shared lists.
- Right stream type: send newsletters, promotions and other bulk mail on a marketing stream. Transactional streams are for mail triggered by a user's own action or account, such as confirmations, password resets, receipts and alerts.
- Honest sender: use a From address at your own verified domain, and don't hide or fake who the mail is from.
- Easy unsubscribe: honor unsubscribes. Marketing streams handle them for you. Don't send to suppressed addresses through other means.
- Prohibited content: no malware, phishing, fraud, or illegal content.
Violations can lead to the stream being paused or banned and, in serious cases, to the account being closed.
Data retention
| Data | Kept for |
|---|---|
| Messages in the sending log and their events | 30 days |
| Webhook delivery records | 30 days |
| Suppression entries | Until removed |
To keep records longer, store them yourself using webhooks or the messages API.