LV.1
EXP 0/1000
โ—€ Playbook index
NO.17

GitHub Codespaces

In one sentence

Codespaces is a disposable development environment hosted in the cloud by GitHub.

Open a browser and a full development environment with VS Code spins up in seconds. The first choice when you want to let AI try something out.

Features

A Docker container in the cloud, where 1 session = 1 codespace. Configuration is codified in devcontainer.json inside the repo.

How to connect (from anywhere)

  • ๐ŸŒ Browser โ€” Open directly from github.com with a single URL
  • ๐Ÿ’ป VS Code Desktop โ€” Remote connection from local VS Code
  • ๐Ÿ–ฅ๏ธ GitHub CLI (gh codespace ssh) โ€” Connect directly from the terminal

Container environment

  • ๐Ÿณ Docker-based โ€” Declare base image + features + post-create script in devcontainer.json
  • ๐Ÿ“ฆ Reproducible โ€” Open the same repo and everyone gets the same environment
  • ๐Ÿ”Œ VS Code extensions pre-installed โ€” Copilot, MCP servers, linters can all be pre-configured
  • โšก Machine size โ€” Switchable on demand (up to 32 cores / 128 GB RAM / 128 GB storage)

When you want to run heavy builds or AI agents in parallel, choose a larger size. You can also change the size after startup.

Benefits

  • ๐Ÿš€ Zero setup time โ€” Just press โ€œOpen in Codespacesโ€
  • ๐Ÿ‘ฅ Same environment for everyone โ€” โ€œIt works on my machineโ€ disappears
  • ๐ŸŒ Access from anywhere โ€” All you need is a browser
  • ๐Ÿ’ช Works on low-spec machines โ€” Even an old PC at hand can use up to 32 cores / 128 GB RAM on the cloud side
  • ๐Ÿ›ก๏ธ Safely isolate AI โ€” If something breaks, throw it away and start fresh
  • ๐Ÿค– AI tools pre-installed โ€” Pre-equipped with Copilot CLI, MCP servers, and Skills