Skip to content
Miloš Vasić

// tier: serverfactory-tertiary · order 29

Mail Server Factory productionlicense: UNVERIFIED

Kotlin (2.0.21)ShellDockerQEMUJSON (declarative config)Gradle 8.14.3 / Java 17CaffeineG1GC-tuned JVMPrometheus-compatible metricsSieve

Source

Mail Server Factory — JSON → Docker mail stack Installed Docker mail stack Connection-type fan-out Security: AES-256-GCM · firewall ports · TLS/HSTS · RBAC · audit log JSON config declarative recipe Factory engine Kotlin 2.0 · JVM G1GC Postfix SMTP Dovecot IMAP / POP TLS / HSTS certificates Sieve filtering Connection targets local · SSH · Docker · K8s · AWS SSM · Azure · GCP · Libvirt
// architecture

Run your mail server like the boss — describe it in JSON, deploy it anywhere.

A Kotlin/Shell tool that turns a JSON description into a fully installed, Dockerized mail server. It supports 12 connection types (SSH, Docker, Kubernetes, AWS SSM, Azure, GCP, Libvirt, and more), a full security framework, 25 Linux distributions, and ships with 439 passing tests.

Mail Server Factory is a production-ready, automated mail-server provisioning tool. The user writes a simple JSON configuration; the Factory interprets it and performs all installations and initializations on the target OS, deploying a Docker-based, loosely-coupled mail stack across 12 connection types.

Standing up a real, secure mail server is one of the classic rites of passage in systems administration — and one of the most reliably miserable. Postfix, Dovecot, TLS certificates, DNS records, firewall rules, and per-distribution quirks all have to line up perfectly, and a single wrong directive means silently bounced mail or an open relay. Mail Server Factory takes that entire body of hard-won, error-prone expertise and captures it in software. Instead of hand-configuring each piece across an unfamiliar OS, the end-user writes the desired outcome as a simple JSON document; the Factory reads that JSON and executes the exact installation and initialization steps required on the target operating system, standing up a mail stack that runs on Docker with every component loosely coupled — a design choice that keeps the stack horizontally scalable and lets any single component be upgraded or replaced in isolation. And it is deliberately reach-agnostic: 12 connection types let the identical tool and identical JSON target a local machine, a remote host over SSH, a Docker or Kubernetes runtime, cloud instances via AWS SSM / Azure Serial Console / GCP OS Login, or VMs via Libvirt — the same declarative description, provisioned wherever you point it. It supports 25 Linux distributions across Western (Ubuntu, Debian, CentOS, Fedora, AlmaLinux, Rocky, openSUSE), Russian (ALT, Astra, ROSA), and Chinese (openEuler, openKylin, Deepin) families, with unattended installation via preseed/kickstart/cloud-init/autoyast and QEMU-based VM automation for testing. Enterprise features are extensive: AES-256-GCM encryption, enforced password and SSH-key policies, automated firewall configuration for mail ports (25/587/465/993/995), TLS/SSL with certificate validation and HSTS, audit logging, and RBAC. Operational features include JVM tuning (G1GC), Caffeine caching, connection pooling, Prometheus-compatible metrics, structured logging, hot-reloading configuration, and secrets management. The project reports 439 tests at 100% pass and a clean SonarQube quality gate. It is the flagship of the Server-Factory org.

The itch we had to scratch

Setting up a secure, production mail server is notoriously error-prone and OS-specific. Mail Server Factory captures that expertise in a declarative JSON model plus an execution engine, so a correct, secured, Dockerized mail stack can be reproduced on any supported target without manual step-by-step work.

The leap it delivers

It demotes mail-server provisioning from a specialist, days-long, get-it-exactly-right ordeal to an act of writing configuration — then makes that act portable across 12 connection types and 25 Linux distributions and hardened by enterprise security defaults out of the box. The result is reproducible and *verifiable*: the same JSON yields the same secured stack every time, and the project's reported 439 passing tests and clean SonarQube gate mean the engine doing the work is itself held to account rather than trusted on reputation.

What's genuinely new

The tough parts — and the fixes

How it's built

Note: GitHub marks the repository as a fork within the Server-Factory org. Predates the AI product line; presented as a mature DevOps/provisioning flagship, not an AI utility.