Development

The desktop editor,on a desktop you rent

Visual Studio Code on an Ubuntu desktop of your own.The application itself, not a browser build of it.

VS Code 1.137.02 cores · 4 GBOpens in a browser tab

hub://vscode

$ dxflow workflow create --identity vscode hub://vscode --start --link

01Pick a machinecores, memory and a rate per hour

02Create the workflowpulled from the hub, once

03Run it, then stopstopped when the work is done

The application

The real editor,not a web build

The desktop application, with Miniconda and Copilot alongside.

The desktop build

The application itself, not a browser port of it, on a real desktop.

Miniconda included

conda sits on the PATH, so Python environments work from a terminal.

Extensions persist

The profile and the extensions directory both sit on the volume too.

How you run it

Every defaultis one override away

The session reads its settings from the environment, so you set them on the start line.

VNC_PASSWORDWhat the tab asks for. It ships as dxflow, and everyone reading this knows that.

PANEL and TASKBARBoth hidden, so the editor has the whole screen. Show either one to get the desktop back.

AUDIOOff here. Turn it on and the desktop sound is streamed beside the picture.

start it your way

$ dxflow workflow start vscode --override env.app.VNC_PASSWORD=something-long

$ dxflow workflow start vscode --override env.app.TASKBAR=show --link

The session

The editorfills the screen

Streamed to your browser, with VS Code already maximized.

6082Browserin the tab, at /vnc.html

5901VNCin a native client

/volumeStoragesettings, extensions, code

Open it from anywhereStart with --link and the session comes back on an HTTPS address.

Set a password firstVNC_PASSWORD defaults to dxflow, and everyone can read this page.

Keep the session yoursIt runs without its sandbox here, so do not pass the link around.

Good to know

It runs as root,so keep it to yourself

Inside the container there is one user, and VS Code drops its sandbox to live with that.

No sandbox

The editor starts with --no-sandbox because root cannot use one. A strong VNC_PASSWORD is what stands in for it.

Miniconda is on the PATH

Installed at /opt/miniconda, so conda builds Python environments straight from the integrated terminal.

Extensions persist

Copilot installs itself on the first launch, into the extensions directory on the volume. It is still there on the next start.

The image

Pulled once,then it stays

VS Code Server arrives as one image. This is what comes down the first time, and what the disk should have free for it.

1.1Gamd64compressed, the way the registry counts it

1.0Garm64compressed, the way the registry counts it

20GOn diskunpacked, with room to work beside it

ghcr.io/dxflow-ai/vscode:latestPublished from the hub, pulled on the first start and kept for the ones after it.

Either architectureamd64 and arm64 are both published, and the machine pulls the one it runs.

What it asks for

What it wants,and what it needs

The definition asks for 2 cores and 4 GB. The image comes up on less than that, and a start given --fit trims the ask to whatever the machine actually has.

2 cores · 4 GBAsks forwhat the definition writes down

2 cores · 2 GBRuns onthe least the image comes up on

Not neededGPUit works on the cores alone

The ask is not the floorThe definition writes down what suits the work. The image itself starts on less, which is what the second figure is.

--fit caps it to the hostA start given --fit trims each step to what the machine actually has, for that start alone. The definition is never rewritten.

Machines that fit it

VS Code Server asks for 2 cores and 4 GB. Cheapest first.

E2 Medium

$0.0419/ hour2 cores · 4 GBStart this machine

T3 Medium

$0.052/ hour2 cores · 4 GBStart this machine

B2s

$0.052/ hour2 cores · 4 GBStart this machine

Run VS Code Server on your own machinePick a machine that meets it, and it opens about a minute after you ask.

Start VS Code Server Browse workflows