GitHub Codespaces Alternative: Code on the Computer You Own

GitHub Codespaces is a good product, and the bill makes sense once you read the meter. But if most of your coding happens on one or two projects, and you already own a capable computer, you may be paying by the hour for a machine that sits idle on your desk. This post walks through how Codespaces billing actually works, does the math for a normal month, and explains when running VS Code on your own computer is the better deal, and when it isn't.

How Codespaces billing works

Codespaces charges for two things: compute while a codespace is running, and storage for as long as it exists. The numbers below come from GitHub's billing docs and the Codespaces product page.

A few details matter more than the hourly rate:

• Free quota is counted in core hours. A personal GitHub Free account gets 120 core hours and 15 GB-month of storage each month. GitHub Pro gets 180 core hours and 20 GB-month. On a 2-core machine, 120 core hours is 60 hours of actual use. On a 4-core machine it's 30.
• Storage accrues while a codespace is stopped. Compute stops when the codespace shuts down. Storage keeps counting until you delete it.
• Organizations get no free quota. If your company pays through a GitHub Team or Enterprise organization, every hour is billed from the first one.

None of this is hidden. It's just easy to forget that a 4-core machine burns through the free tier twice as fast as the default.

The cost math for a normal month

Say you code about four hours a day, 22 working days a month. That's 88 hours. Here's what the compute side looks like on a personal Free account, before storage:

Bill the same 88 hours through an organization, with no free quota, and a 4-core seat runs about $32 a month per developer in compute alone. Add storage: a few codespaces at 10 GB each, kept around between sessions, push you past the free 15 GB quickly.

Compare that with a flat plan. Coder IDE's Free plan is $0 for one computer. Pro is $12 a month for unlimited computers. You already paid for the hardware, so the hourly meter just goes away. Full details are on the pricing page.

The honest version: if you use a 2-core codespace a few hours a week, Codespaces is probably free for you, and you should keep using it.

A cloud VM versus your own machine

Price is only part of it. The bigger difference is where the work happens.

Your machine is usually faster. A modern laptop or desktop typically has more cores, more memory and a faster disk than a 4-core codespace. If you have a GPU for local models, game engines or ML work, a codespace can't use it.

Your files and setup are already there. Your dotfiles, SSH keys, database with real test data, Docker images you already pulled, and the half-finished side project that isn't in a repo yet. With Codespaces, all of that has to be recreated in a devcontainer or copied up.

Your code stays on your machine. With Coder IDE, the editor runs on your own computer and your browser connects to it. Traffic passes through a relay that forwards it; the code itself never moves to someone else's disk. On the same home network, the connection goes direct and only falls back to the relay when it needs to. More on that in how it works and security.

The tradeoff: your machine has to be on. A codespace runs whether your laptop is open or not. Coder IDE needs the computer you're connecting to to be awake and online. A desktop that stays on, or a laptop set not to sleep on power, solves this for most people.

When Codespaces is still the right choice

Be fair to it. Codespaces wins in several common cases:

• Ephemeral, per-branch environments. Spinning up a clean environment for every pull request, then throwing it away, is exactly what Codespaces is designed for.
• Onboarding. A new contributor can open a repo with a working devcontainer and be productive in minutes, without installing anything.
• Open source drive-by fixes. You want to fix one typo in a project you'll never touch again. A codespace is cleaner than cloning it locally.
• Locked-down laptops. If your work laptop can't run the toolchain, a cloud VM may be the only option.
• You don't own a capable computer. If your only device is a tablet, you need a machine somewhere. Codespaces is one; a spare desktop at home is another.

Many people use both: Codespaces for throwaway environments, their own machine for the projects they work on every day.

Moving your daily workflow to your own computer

If you decide to try it, the setup is short:

1. Install once. Run the install command shown on the Coder IDE site on the computer you want to code on. Windows, macOS, Linux and WSL all work.
2. Approve the machine. Click Sign in in the desktop app and confirm the code from a browser you're already signed in to. Every new machine needs your approval, and the approval screen shows which computer is asking.
3. Open it anywhere. The machine appears on your dashboard. Tap Connect (or Open, if the editor is already running) and you're in VS Code from a browser, an installed web app, or the Android app.

What you get is Microsoft's own VS Code server running on your machine, with your extensions, settings, keybindings and integrated terminal. It's not a lookalike editor, so the extensions you rely on behave the way they do at your desk. If you're curious why that matters, code-server vs VS Code serve-web vs Remote Tunnels covers the extension differences in detail.

There are no ports to forward, no VPN to set up and no certificates to manage. It works behind NAT and firewalls.

Things that replace Codespaces features

• Multiple environments: open each folder as its own project on the same machine and switch between them.
• Port forwarding: dev servers you start show up in the Ports panel. Turn on Allow preview for a port and it gets a private link that only opens for you; the Preview button opens the running app beside your code or in its own tab. See open your localhost dev server on your phone.
• Idle shutdown: turn on auto-suspend, and idle editors shut down after a timeout you choose. It waits for running Claude Code tasks before suspending.
• Devcontainers: they still work locally with VS Code's Dev Containers support if your project already has one.

Teams: shared devices and SSO

Codespaces is often adopted at the org level so that nobody has to maintain a build machine. If your team already has strong workstations, or a shared build box under a desk, Coder IDE's Team plan ($49) adds seats, shared devices and SSO. A shared device lets several teammates open the same powerful machine without passing credentials around, and removing a device from the dashboard stops its token immediately.

If your organization needs centrally managed, policy-enforced, disposable environments for hundreds of developers, a cloud platform is a better fit. Coder IDE is built around computers people already own.

The short version

Codespaces bills you for a cloud computer by the hour, plus storage while it sits idle. That's a fair deal for disposable environments and for people without a capable machine. For your everyday projects, on a computer you already own, a flat plan and your own hardware are usually cheaper and faster, and your files stay where they are.

If your setup lives on more than one computer, or you want to reach it from a tablet, your computer is the best cloud dev environment you own goes further into that approach.

Start free. The free plan covers one computer.