Skip to main content
Sometimes you want more than a single send but less than a forever-running automation. A product launch is the classic case. An announcement, a follow-up two days later, but only to people who didn’t open the first one. That’s a flow, not an email. It just needs to run once, against an audience you choose, at a moment you choose. A manual-audience automation does this. You build a normal flow: sends, waits, filters, splits. Instead of listening for an event forever, it runs your saved audience through the graph when you launch it. Treat it as a one-off, not a live automation. This recipe builds a two-stage launch sequence. The same pattern fits event invites, seasonal promos, and any send whose follow-up depends on what the first email did.
Before you start. You need a verified sending domain and at least one saved audience.

1. Save the Audience

Go to Audience in the sidebar and build the segment you want to hit. Filter by plan, activity, custom fields, or whatever defines the group. Click Create new audience in the bar at the bottom of the page, and give it a clear name. The automation’s trigger node shows this audience, its filter query, and the live contact count. That count includes unsubscribed members, and the run skips them. A dryRun returns the eligible recipient count. Filters in the flow, and contacts who unsubscribe before their send, can still lower the number delivered.

2. Generate the Flow

Open Home and describe the sequence, making the one-off intent explicit:
Brew opens the flow on its own canvas with the trigger, both emails, the wait, and the filter in place. If the trigger isn’t set to your audience yet, click the trigger node and use Change trigger.

3. Review the Emails

Click each Send Email node to review it. Refine in chat (“Tighten the follow-up subject line”) or edit the email directly. Personalize with {{ firstName | there }} as usual. The filter node is what makes this a flow instead of two separate sends. Contacts who opened email 1 exit quietly, and only non-openers get the nudge.

4. Test the Run

  1. On the canvas, open Automation options → Test automation. Enter your own address under Send test to. Brew skips waits and delivers test emails only to this address.
  2. Leave Control test scenario unchecked and click Run test. With no open event, the filter lets the follow-up through. You should receive both emails.
  3. Open Test automation again and enable Control test scenario. In the Engagement step that watches email 1, click Add event and choose Opened. Leave Milliseconds after this window opens at 0.
  4. Click Run test. On the opener path, you should receive the announcement only.

5. Run It

When it looks right, start the run with Run now, Schedule, or Gradual send. There’s no publish step: a manual-audience automation has nothing to leave running in the background. While contacts are moving through the flow, the automation shows Sending in the Automations list. When the run completes, it shows Sent. Gradual send delivers each send step in hourly or daily batches instead of to everyone at once. Use it when the domain is newly verified or has been sending at low volume. Send Options has the limits on a scheduled or gradual run, and what happens when your organization requires send approval. Intelligent send isn’t offered here, since timing inside a flow comes from its Wait steps. A run already in flight can be paused, resumed, or canceled; see From an Agent or Code.
Every send in the run counts toward your monthly send limit. Check the contact count on the trigger node before launching. From the API, dryRun: true previews the eligible recipient count without sending, and scheduledAt launches the run later.

6. Reuse It Next Time

The flow stays in your Automations list after the run. For the next launch, open it and swap the email content in chat (“Update this for the [new feature] launch”). Point Change trigger at a different audience if needed, and run it again. Each run is tracked separately, so the history of past launches stays intact.

From an Agent or Code

Over MCP, create_email makes each design first. save_automation then saves the flow, with a manual-audience trigger pointed at the saved audienceId and each send node pinned to one design. test_automation with a testRecipient runs it to one address. To test the path where openers exit, add scenario with an open on the filter node. run_automation launches it: dryRun: true returns the recipient count without sending, scheduledAt launches later, and gradualSend sets the batches. On OAuth and organization connections a real launch returns a confirmation request first. Repeat the call with confirmed: true and the returned confirmation_id after you approve. See Confirmation. Over the API the launch is POST /v1/automations/{automationId}/run with the same fields and an Idempotency-Key. A run in flight is controlled with pause_audience_run, resume_audience_run, and cancel_audience_run, or POST /v1/automations/audience-runs/{audienceRunId}/pause, /resume, and /cancel, each with the run’s audienceRunId.

Need Help?

Our team is ready to support you at every step of your journey with Brew. Choose the option that works best for you:

Search Documentation

Type in the “Ask any question” search bar at the top left to instantly find relevant documentation pages.

Ask ChatGPT or Claude

Click here to open the Copy page menu at the top of this page, then send the page to ChatGPT, Claude, or your editor.