Govern AI Email Actions Inside Your Make Scenarios

Use MultiMail's REST API from Make's HTTP module to send, read, and approve emails with full audit trails — without writing a backend.


Make (formerly Integromat) lets teams build complex automation workflows by connecting SaaS tools through a visual scenario editor. When those scenarios involve sending email — especially email triggered by AI logic, CRM events, or webhook payloads — you need more than a simple SMTP relay. You need audit trails, delivery tracking, and the ability to require human approval before a message goes out.

MultiMail integrates with Make through its REST API via Make's native HTTP module. Any scenario can call MultiMail endpoints to send email, check an inbox, read a thread, tag messages, or trigger the `POST /v1/oversight/decide` approval flow — all with Bearer token authentication and structured JSON responses that map cleanly to Make's data mapping layer.

The combination is especially useful for teams running AI-assisted outreach, support escalations, or approval-gated notifications. Make handles the trigger logic and routing; MultiMail handles delivery governance, oversight mode enforcement, and the human-in-the-loop approval queue.

Built for Make (Integromat) developers

Approval gates before every AI-generated send

Set a mailbox to `gated_send` or `gated_all` oversight mode. Any HTTP POST to `POST /v1/mailboxes/{mailbox_id}/send` from your Make scenario queues the message for human approval instead of sending immediately. Your scenario continues; the email waits. MultiMail fires a webhook to a separate Make trigger when the approver acts.

Structured inbox access for downstream modules

The `GET /v1/mailboxes/{mailbox_id}/emails` (list inbox) and `GET /v1/mailboxes/{mailbox_id}/emails/{email_id}` (read one) endpoints return clean JSON — sender, subject, markdown body, thread ID, tags. Make's data mapper can route fields directly into Airtable rows, Slack messages, or CRM updates without regex parsing or fragile HTML scraping.

Per-mailbox audit log for compliance

Every send, read, approval, and rejection is logged against the mailbox. For teams subject to CAN-SPAM, GDPR Article 5 accountability requirements, or internal SOC 2 controls, this log is the audit trail that proves governed operation — not just that email was sent.

Webhook triggers for inbound email

MultiMail fires a `message.received` webhook when a message arrives at your mailbox. Drop a Webhooks module at the start of a Make scenario to parse the payload and branch on sender, subject, or tags — no polling loop needed.

Oversight mode you can change per scenario

A mailbox used for high-volume notifications can run `monitored` while the same tenant's customer-facing mailbox stays `gated_send`. Switch modes via the API when a scenario's risk profile changes, without redeploying anything.


Try it with your agent

No code, no dashboard. Paste this to your AI agent — it connects MultiMail, creates an inbox, and builds the flow for you.

1. Get MultiMail ready first: read https://multimail.dev/llms.txt, connect the MultiMail MCP server to this agent, create a free inbox, and set up a verified sender so the agent can compose, send, and schedule email from an approved identity. 2. In Make (Integromat), create or open your AI Agent, then add MultiMail through Make’s built-in HTTP app/module as the tool the agent can call for email actions. Use only the MultiMail details provided by llms.txt and the connected MCP server; do not invent endpoints, headers, or webhook flows. 3. Give the Make agent exactly three email capabilities: check the MultiMail inbox for new messages, draft replies for developer review, and send or schedule approved messages from the verified sender. 4. Send a test email to the new MultiMail inbox, have the Make agent check the inbox, draft a short reply, and show the draft back to me before taking any send action. 5. Run MultiMail in gated_send mode for this workflow: every outbound email must pause for my review and approval before it sends, while keeping the agent’s email actions auditable. Do not switch to monitored or autonomous unless I explicitly ask.

Step by step

1

Create a MultiMail account and get your API key

Sign up at multimail.dev and create a mailbox — either a custom domain mailbox or one ending in `@multimail.dev`. Copy your `mm_live_...` API key from the dashboard. If you want approval gates, set the mailbox oversight mode to `gated_send` now.

2

Add an HTTP connection in Make

In your Make scenario, add an HTTP > Make a request module. No official MultiMail connector exists yet, so all calls go through HTTP directly. Store your API key in Make's Connection vault or as a scenario variable — do not hardcode it in module configuration.

3

Register a webhook URL for inbound events

Create a new scenario starting with the Webhooks > Custom webhook module. Copy the generated Make webhook URL. Then POST to `https://api.multimail.dev/v1/webhooks` with your Make URL and the events array `["email.received", "email.delivered"]`. MultiMail will now push events to your scenario.

4

Map MultiMail response fields to downstream modules

Run your scenario once with a real email to let Make detect the webhook payload structure. Make will infer the field schema from the first real payload, making `email_id`, `thread_id`, `from.address`, `tags`, and `subject` available as mappable variables for all downstream modules including Routers, Filters, and data store writes.

5

Test with MultiMail test keys before going live

Replace `mm_live_` with `mm_test_` keys during scenario development. Test-mode sends are recorded and inspectable in the dashboard but never delivered to real recipients. Switch to live keys only after the full scenario flow — including the approval branch — has been validated end to end.


Common questions

Is there a native Make module for MultiMail?
Not yet. Use Make's HTTP > Make a request module with the MultiMail REST API directly. All endpoints accept and return JSON, and Make's data mapper handles the response fields without any custom parsing. A native connector is on the MultiMail roadmap.
How do I avoid accidentally sending emails during scenario testing?
Use a `mm_test_...` API key. Test-mode requests go through the full API validation path and return realistic responses, but no email is ever delivered. You can inspect test-mode sends in the MultiMail dashboard under the test environment tab.
What happens if a Make scenario errors after the send endpoint is called?
If the HTTP module call to the send endpoint returned a 200 before the error, the message is already queued (or sent, depending on oversight mode). Make's error handling does not roll back completed API calls. For sensitive outbound sends, use a `gated_send` mailbox so messages sit in the approval queue until you explicitly confirm them — giving you a manual cancel window.
Can I use Make to build the human approval UI?
Yes. When MultiMail fires an `approval.requested` webhook, your Make scenario can POST the pending message details to a Slack channel with Approve/Reject buttons, or create a row in Airtable with a status column your team updates. When the approver acts, call `https://api.multimail.dev/v1/oversight/pending` to fetch and then POST `/v1/oversight/decide` with `{ "email_id": "...", "action": "approve" }` (or `"reject"`) to complete the flow.
How do I handle rate limits from MultiMail inside a Make scenario?
MultiMail returns HTTP 429 with a `Retry-After` header when rate limits are hit. Configure the HTTP module's error handling to retry on 429 — Make supports up to 5 automatic retries with configurable delay. For high-volume scenarios processing many emails in parallel, use Make's Flow Control > Sleep module to throttle between HTTP calls.
Can a single Make scenario manage multiple MultiMail mailboxes?
Yes. Pass the `mailbox_id` field dynamically using Make's data mapper — map it from the incoming webhook payload or from an upstream data store lookup. Each HTTP module call is independent, so a Router can branch requests to different mailboxes based on any condition in the scenario data.

Explore more

The only agent email with a verifiable sender

Email infrastructure built for AI agents. Verifiable identity, graduated oversight, and a hosted MCP server. Formally verified in Lean 4.