Lockally

      Use your API key in the examples

      Replaces lk_live_… in the code samples on this site so you can copy and run them. It is stored in this browser only and is never sent anywhere, including to Ask AI.

      It will be visible on the page. Use a test key if you are sharing your screen or taking screenshots.

      High-volume sending

      “High volume” on this platform means anything large enough to meet a cap, and the caps arrive sooner than most people plan for. This page is the practical sequence.

      Before the first send

      Work out which cap binds first. Three apply at once, and the smallest one is your real ceiling:

      Where it comes from
      Messages per minute Your plan. From 1 on Free up to 300.
      Messages per day Your domain’s warm-up day, if it is under 30 days old.
      Messages per period Your plan’s included allowance, then prepaid units.

      A workspace on a plan allowing 60 a minute, with a domain on day 6 of its ramp, can send 500 that day and no more. The per-minute number is irrelevant to that constraint.

      Start the domain warming before you need it. Verification starts the clock. A domain verified the week before a launch is on day 7 at launch, which is a 500 message ceiling. Verify a month ahead and the ramp is finished.

      Use a subdomain for bulk. news.yourcompany.com, separate from your transactional domain and from your root. This is much harder to retrofit after a bad campaign than to do now. See Domain reputation.

      Authenticate and confirm alignment. Send yourself one message and read the headers before sending forty thousand. See Email authentication.

      Building the send

      Use batch, and size it for retries. POST /v1/send/batch takes up to 500 messages per call, each with its own recipient and variables. Results come back per entry keyed by index, and entries succeed and fail independently, so a failure in one does not cost you the other 499.

      One message per recipient. A single send to 500 addresses renders once and gives you one bounce for the lot, and everyone sees everyone. See Personalization data.

      Give every message an idempotency key, derived from something stable like the recipient and the campaign, not a random value. That is what makes a retry after a timeout safe.

      Set "unsubscribe": true. On bulk sending this is not optional in practice: it adds the one-click header providers now expect, and it is what makes the send count as broadcast rather than transactional. See Unsubscribe management.

      Pacing

      Spread across the day. A daily cap spent in the first minute leaves the rest of the day rejected. Divide your volume by the hours you are willing to use and pace to that.

      Honour Retry-After. Every pacing 429 carries one.

      Spread across providers where you can. There is no per-recipient-domain throttling here, so a batch aimed entirely at one provider arrives as fast as your cap allows, which is exactly the shape that gets throttled at their end.

      Afterwards

      Read the outcomes within hours, not days. Over decided messages only: bounced / (bounced + delivered). See Monitoring.

      Reconcile the suppression list against your own database. Suppression stops the next send, it does not clean your list, and your bounce rate is computed from what actually went out.

      Expect unconfirmed. It means the outcome was never learned, not that anything failed, and it is common. Do not resend on it.

      The sequence, short

      1. Verify the bulk subdomain a month early.
      2. Confirm DKIM and DMARC alignment on a test message.
      3. Work out the binding cap for launch day.
      4. Batch, one message per recipient, idempotency keys, unsubscribe: true.
      5. Pace across the day and honour Retry-After.
      6. Check the bounce rate after the first few thousand.
      7. Reconcile suppressions back into your own data.

      What the platform will not do for you

      It will not verify your list, warn you that a list looks bad before you send it, pace your sending on your behalf, or tell you a message landed in spam. Those are the parts you own, and on a large send they are the parts that decide the outcome.