Deployment guide

Deploy Secure Message in an afternoon

Stand up the gateway on your own infrastructure: one Debian host, filtering inbound mail and protecting outbound within an afternoon. Here’s the full path, start to cutover.

Before you start

What you’ll need

A host to run it

A dedicated Debian 13 server or VM, 2 vCPU and 4 GB RAM to start, on your own infrastructure. Ports 25, 80 and 443 reachable inbound, and outbound 25 open.

A hostname and reverse DNS

A public FQDN pointing at the host, plus a PTR record matching it. Both are needed before a certificate can be issued or your mail will be trusted.

DNS access

Ability to edit MX, SPF, DKIM, and DMARC records for the domain(s) you want to protect.

Step by step

From license to live in five steps

1

Prepare the host

Install Debian 13 on a dedicated server or VM, set a static address and FQDN, and publish the A record plus matching reverse DNS.

2

Install the gateway

Download the release and run sudo ./install.sh. One prompt, then create the first administrator and complete the five-step setup wizard in the browser.

3

Activate

The gateway shows its node ID under System → License. Enter it with the activation code from your email to receive the licence key, then apply it. Or evaluate on the Community tier with no key at all.

4

Configure policy

Set inbound filtering thresholds and outbound encryption rules, generate DKIM keys, and issue the TLS certificate — all from the console.

5

Verify, then cut over

Run the built-in preflight, send a real message end to end, then point MX at the gateway and set it as your outbound smart host. No mailbox migration, no client software.

Deployment model

Self-hosted, with AI on your terms

You run it

One Debian host. Nothing in the mail path depends on us being reachable.

  • Runs entirely inside your perimeter
  • Your keys, your storage, your retention
  • No cloud component in the mail path
  • Data-residency and sovereignty friendly

AI stays where you put it

Point the classifier at a hosted model, or at an endpoint inside your own network.

  • Filtering stack runs fully on-host
  • Optional AI can use a hosted model…
  • …or inference on your own server
  • Message content need never leave your network

Exact commands, ports, defaults and verification steps are in the documentation. Need a hand? support@secure-message.io.

Ready to deploy?

Start free on the Community tier, or talk to us about a rollout.