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.

      Sending rate management

      A send can be refused by four different limits, all of them 429, all with different meanings and different correct responses. Telling them apart by the title is the whole job.

      The four

      Title Limit Resets Retry-After
      Rate cap exceeded Your workspace’s messages per minute Rolling 60 seconds 60
      Domain warm-up cap reached That domain’s daily ramp UTC midnight seconds until midnight
      Warm-up cap reached The platform’s daily ramp UTC midnight seconds until midnight
      Monthly allowance reached / Out of send units Your plan’s included volume Billing period none

      The first three are pacing problems. The last is a billing one and will not clear by waiting a minute.

      They are checked in that order, so the per-minute cap is what you meet first.

      The per-minute rate cap

      Set by your plan, measured over a rolling 60 seconds:

      {
        "status": 429,
        "title": "Rate cap exceeded",
        "detail": "Your tenant is capped at 10 sends/min. Try again in ≤60s."
      }

      Real values across plans today run from 1 per minute on Free up to 300 on the largest. The great majority of workspaces are on the lowest tier, so if you are testing on a Free workspace, one send per minute is the behaviour to expect and not a fault.

      Pacing a client properly

      Honour Retry-After. Every 429 here carries one where waiting helps. Guessing instead is how a client turns a one-minute pause into a retry storm.

      Spread, do not burst. A daily cap consumed in the first minute leaves the rest of the day rejected. If you have 2,500 messages and a 2,500 daily cap, send them across the day rather than at 00:01.

      Use batch for volume. POST /v1/send/batch takes up to 500 messages in one call, and each one still counts individually against every limit above. It saves round trips, not allowance.

      Queue on your side. The limits are a property of your plan and your domain’s age, not of the request. A client that treats 429 as an error to surface, rather than a signal to slow down, will lose mail.

      async function sendPaced(body, key) {
        for (let attempt = 0; attempt < 5; attempt++) {
          const res = await send(body, key);           // same idempotency key each time
          if (res.status !== 429) return res;
          const wait = Number(res.headers.get('retry-after') ?? 60);
          await new Promise((r) => setTimeout(r, wait * 1000));
        }
        throw new Error('still rate limited after 5 attempts');
      }

      Reuse the idempotency key across attempts so a retry after an ambiguous failure cannot send twice. See Idempotency.

      Per-recipient pacing

      There is none. The platform does not throttle per receiving domain, so a batch aimed entirely at one provider goes out as fast as your own cap allows. Large volumes at a single provider are better spread by you.

      Platform-wide outbound pacing exists as a setting and is currently switched off, so the only pacing in force is the per-workspace cap.

      Raising a limit

      Rate caps come from your plan, so the usual answer is the plan. If your volume genuinely needs more than your tier allows, that is a support conversation, and the useful things to bring are your sustained volume, your bounce rate and how the sending is spread through the day.