Shipown is a cross-platform Electron desktop application that deploys and operates web applications on Linux servers you own. It gives you the workflow of a managed deployment platform — connect a source, build, get a live URL, roll back — while the servers and the data stay yours.

How is Shipown different from Vercel or Netlify?

Managed platforms run your app on their runtime and their account. Shipown deploys to a server you already rent or own, over plain SSH, and installs no agent to do it. You keep the smooth workflow without handing over the infrastructure.

What can it deploy today?

Three runtimes. Node applications run by PM2 — Next.js, Nuxt, Remix, SvelteKit, NestJS, Fastify and plain Express. Static sites served by Nginx, including Vite with React or Vue, Astro and plain HTML. And Docker, from a Dockerfile or a compose file, which covers any language that ships in a container. Python, Go, Rust, PHP, Ruby, Java and .NET have no native runtime here; deploy them in a container.

Does Shipown need Docker?

No. Node applications run directly under PM2 and static sites are served by Nginx, on a server with no Docker at all. If you do use Docker, Shipown deploys to it as well — a Dockerfile or a compose file, with releases, health checks and rollback — and manages the containers already on the server: every container inspected, published and internal ports mapped, CPU and memory gauges, crash loops called out, start, stop, restart, a shell inside the container, and its logs. It can also install Docker on a server that has none.

Where does it deploy to?

A Linux server you can reach over SSH — a VPS, a cloud VM, or a box in your office. Ubuntu, Debian, the RHEL family and Alpine are handled. Your own computer is where a project comes from, not somewhere Shipown deploys to.

Are my SSH keys or environment variables uploaded anywhere?

No. SSH private keys, passwords, passphrases, environment variable values, source code and project files never leave your machine for licensing or telemetry. Credentials are held by the Electron main process and are never exposed to the interface. The one exception is one you opt into: the AI terminal sends its conversation to the AI provider you pick.

Where does the AI terminal send my server’s data?

Only to the provider you choose (DeepSeek, Anthropic Claude, or OpenAI), and only while you actively interact with it. The AI operator sends your prompt and the terminal outputs of the commands it runs under your own encrypted API key. Shipown does not intercept or store this telemetry. Destructive commands (reboots, rm -rf, firewall changes, stopping SSH) are never executed automatically without your explicit confirmation.

Can I roll a bad deployment back?

Yes. Every deploy lands in its own release directory, and the live pointer moves only after the build and the health check pass. A deploy that fails restores the previous release by itself, and you can switch back to an earlier release later from the app.

Does it deploy automatically when I push?

No. Shipown tells you when a branch has moved ahead of what a server is running, and how many commits are waiting, but a person still presses deploy. Automatic deployment on push is not part of it.

Is Shipown open source?

No. Shipown is a commercial product and its source is not published. What runs on your servers is plain SSH and shell — nothing is installed to phone home — but the desktop application itself is licensed, not open.

Is Shipown free, and how do I get access?

Every plan starts with a free 30-day trial and no card is asked for until you buy. What you get is a licence key with an expiry and a device limit — typically two machines. Activation is checked by the licence server rather than by the app, so a key can be extended, replaced, or have a device slot freed without reinstalling anything.

Which platforms does the desktop app run on?

Windows, as an .exe installer, and Linux, as a .deb for Debian and Ubuntu. Both come from the same download server the app updates from, so the file on the download page and the file an update fetches are the same file. macOS is not published yet.

Docs/Product Modules/Servers and monitoring

Servers and monitoring

Save servers, watch SSH reachability independently of the interface, and open live gauges, top processes, an SSH terminal and apps.

Open in App

Connection lifetime and monitoring

Healthy sessions do not disconnect after a minute of inactivity. The backend monitors enabled, unlocked saved servers even while the window is minimized or hidden in the tray. Idle SSH sockets can close after five minutes without an active command and reopen as needed; socket cleanup never means the server is offline or its session password is forgotten.

Manual Disconnect pauses monitoring and forgets session-only credentials. Reconnect or unlock to resume. Locked servers are not probed, and passwords saved with Remember disabled stay in memory only. Older disconnected profiles require one explicit Reconnect.

SSH outage and recovery alerts

A separate SSH heartbeat runs every 15 seconds for the first ten live servers while visible, or every 60 seconds while hidden and for additional servers. One failed check shows RETRYING. Two consecutive failures show SSH UNREACHABLE and report one incident in the Live Feed, an OS notification and an in-app toast when alerts are enabled. Retries continue with backoff up to five minutes; a successful check reports recovery once.

PM2 or Docker command errors do not declare SSH unreachable. This alert means SSH cannot be reached from your PC; it does not prove that the machine or its apps stopped. Check the PC network, server and firewall. Expensive live PM2 comparisons and Docker streams are limited to ten servers, while heartbeats and background application checks cover additional enabled sessions.

Settings and desktop limits

Server & application alerts controls SSH outage/recovery and background PM2/Docker notifications. Application sweeps have their own 5–180 minute interval, defaulting to 15 minutes; SSH heartbeats run separately. Turning notifications off leaves the Live Feed running.

The desktop app must be running and the PC awake. Tray/minimized monitoring works, but no desktop checks run after quitting or while the PC is asleep/off. Resume starts fresh checks without treating sleep as a server outage. Idle interface time never cancels a deploy/build or clears its session credentials.

← PreviousAI Terminal & Reasoning Next →Docker Containers