There’s a pattern I keep seeing in small engineering teams.

You start with a simple setup. A few servers, a few containers, one databse, SSH and some shell scripts. It works. You move fast.

Then you grow a little. More services, more environments, maybe a second region. Multiple Databases. Distributed Caches. The scripts get longer. The wiki page nobody reads gets a new section. Deployments are a 12-step checklist someone keeps in their head.

So you look at the options. Kubernetes. Nomad. Portainer. Coolify. Each one solves the problem but introduces a new layer of abstraction you now have to understand, maintain, and debug when it breaks.

You traded one kind of complexity for another.

The original pain was real. But the jump from “shell scripts that work” to “platform that works” doesn’t have to go through “platform that requires a platform team.”

There’s a middle ground that most tooling skips right over. Simple enough to run on day one. Solid enough to trust at 3am. Small enough that one person can hold the whole thing in their head.

We went looking for that middle ground. Didn’t find it. So we started building it.

Media