ORKASTORCLOUD

Preview environments

Push code.
Get a live URL.
It cleans itself up.

Connect a repository and every branch gets its own running copy of your app — the frontend, the API and the database together, on one address. Each copy shuts down on a timer, so a test environment never quietly turns into a bill you forgot about.

Early access, by invitation. You deal with the people who built it, not a support queue.

Every branch its own environment. Each one leaves when its time is up.

One staging environment.
Five open pull requests.

Somebody's change is always waiting behind somebody else's, the screenshot in the review is never quite the bug, and "works on my machine" is still how most of it gets settled.

Giving every branch its own running copy fixes that, and teams have known it for years. The reason it stays on the wish list is that the copies pile up: nobody remembers which ones are still needed, so either somebody spends a morning a week clearing them out, or the bill quietly grows. Putting them away by themselves is the part we built first.

Three steps, then it is automatic.

  1. 1

    Connect a repository

    Point Orkastor at a repo and pick a branch. We work out how to build it and show you what we found before anything runs.

  2. 2

    Push

    A push builds the app and starts an environment. You watch the build output as it happens, and you get the address the moment it serves.

  3. 3

    Walk away

    Set how long it should live, or the hours it should be awake. It stops on its own. Nothing to remember, nothing left running over a weekend.

What you get

A whole environment, not a single process. These are the things people ask for first.

Your app and its database, together

Run the frontend, the API and Postgres or Redis side by side. They share one address, so your frontend calls your API the same way it does on your laptop.

Private unless you say otherwise

A new environment is not on the public internet. Share it with a password, lock it to your team's sign-in, limit it to known addresses, or keep it reachable only from inside your own environments.

Yours alone

Your environment does not share space with another customer's. That separation is the part we are least willing to compromise on, and it applies to every environment, not just the paid ones.

An address that works immediately

Every environment gets its own HTTPS address as soon as it starts. No certificates to request, no DNS to edit, nothing to configure.

Awake only when you need it

Give an environment a timer, or a weekday schedule down to fifteen minutes. It sleeps outside those hours and wakes when the window opens.

The output you actually want

Build logs, running logs, environment variables and files, all in one place. When a build fails, the reason is on the screen rather than in an archive.

You should never be surprised by the bill.

Most of the fear around environments like these is not the price. It is not knowing what the price will be. So the product is built to answer that before you spend anything.

  • See the cost before you create it. The Console prices an environment while you are still describing it.
  • Set a hard limit. Choose a ceiling for the month. When it is reached, environments stop rather than keep charging.
  • Pay for time that ran. An environment that is asleep, expired or never woken is not billed for that time.

Early access is open

We are taking on a small number of teams by invitation. Being early is worth something here: you talk to the people building it, and what you ask for goes into the product rather than onto a backlog.

The honest state of it, because you should hear it from us rather than find out later: it runs in one region today, eu-north-1. Everything described on this page is built and running — nothing above is a plan. When something is not ready, it does not go on this page until it is.