01
One command, every stage
portway up detects your stack and provisions preview, staging, and production from the same binary. Same inputs, same output, no per environment snowflakes.
One command deploys
Your deploy should be one command, not a runbook.
Portway reads your repo, builds a reproducible, content addressed image, and ships it to your own cloud in seconds. No YAML sprawl. No dashboard archaeology. No 3 a.m. rollback that needs a wiki page.
macOS, Linux, and CI
Rolls back in one command
No lock in
portway up: from commit to production in a single pass
Why teams switch
Deploys should be boring. Portway makes them boring on purpose.
The tool that got you to launch should not be the thing that wakes you up. Portway trades pipeline glue and dashboard clicks for one predictable command that does the same thing every time.
02
Reproducible or it did not happen
Every build is content addressed. The artifact you ship is the exact artifact you tested, so a rollback is one command and never a rebuild you have to pray over.
03
Your cloud, your keys
Portway deploys into your AWS, GCP, or bare metal. No control plane owns your traffic, no per seat bill, and no vendor stands between your users and your app.
Get started
One line to install. One command to ship.
Drop the installer into your shell, then run Portway from the root of your repo. It figures out the rest and tells you exactly what it is doing while it does it.
Works on macOS, Linux, and any CI runner. No agent to babysit, no daemon left running.
1. Install
$ curl -fsSL https://get.portway.dev | sh
2. Deploy
$ portway up # detect, build, test, ship
The point of view
Fewer moving parts is a feature.
Most deploy tooling grows until it needs a team to run it. Portway keeps the surface small: one binary, one command, artifacts you can trace back to a commit. If a deploy is ever surprising, we consider that a bug in the tool, not a lesson for you to memorize.