Introducing Kaja: push to deploy on servers you already own

There are two decent ways to ship a backend service today, and both cost you something.

You can hand the whole problem to a hosted platform. The developer experience is excellent, and you stop thinking about infrastructure. In exchange, someone else owns the machines, your data lives where they choose, and every container, build minute, and gigabyte shows up on an invoice that grows with your success.

Or you can run it yourself. You keep control and the hardware is already paid for — but now somebody on your team owns YAML, ingress, certificates, secrets, and the pager. That person is usually your best engineer, and this is not what you hired them for.

Kaja is the third option. Connect a server you already run, point a project at a GitHub repo, and push. You get the workflow of a hosted platform on infrastructure that stays yours.

What that looks like

• Push to deploy — Connect a repo once. Every push builds and rolls out. There is no pipeline to wire up, and the build runs on your own server.
• Deployments you can undo — Every release is a versioned revision, with the full history in one timeline. A bad deploy is one click away from being the previous one.
• Projects, not namespaces — Dev, staging, and prod are separate projects with their own apps, config, and secrets. Duplicate one when you need an exact copy, or save it as a blueprint to spawn more later.
• Domains and HTTPS — Every app gets a URL the moment it deploys. Attach your own domain and certificates are issued and renewed for you.
• Managed services — Add observability to a project in one click and bind it to your apps. PostgreSQL, Redis, and Grafana are in the catalog and landing soon.

Your compute is never metered

This is the part we'd most like you to check. Your apps, builds, logs, and metrics all run on hardware you already pay for. That costs us nothing, so we don't charge for it — there is no per-container, per-build-minute, or per-gigabyte-of-logs line item, and there won't be one.

What we charge for is the control plane and the front door: the dashboard across every server you connect, the roles and audit log on top of it, and the gateway that publishes apps from servers with no public IP. Traffic served directly by your own load balancer never passes through us and is never counted.

Nothing dials in, nothing is locked in

You install an agent that only makes outbound connections. No open ports, no exposed API server, no shared kubeconfig. And underneath the console, everything Kaja creates is a standard Kubernetes object you can inspect and keep. If you stop paying us tomorrow, your apps keep running — which is the only version of "no lock-in" worth the words.

Where we are

Kaja is in beta and free while we're there. One server stays free after that. If you have a machine sitting there running k3s and a repo you're tired of deploying by hand, that's exactly the case we built this for — tell us how it goes.