Skip to content

Guides

Welcome to the runbooks. These guides are meant to be concrete, step-by-step instructions—not just high-level ideas.

Each guide explains exactly how to run a specific app or AI agent inside Terrarium. You'll learn:

  • What to install: The exact commands to run inside your LXC container.
  • How to create the container: Step-by-step using either the sleek LXD web UI or the terminal.
  • How to configure it: A humane, interactive path for setting up the software.
  • How to publish it: Whether the app should stay completely private, or be exposed to the internet using Terrarium's user.proxy labels.
  • How to secure it: The best way to handle authentication (like locking a web app behind Single Sign-On).

The Golden Rule

Remember: Everything inside a Terrarium container is completely private by default.

You can run databases, internal APIs, and experimental AI agents without worrying about accidentally exposing them to the public internet. Only the routes you explicitly tag with a user.proxy label will become internet-facing.

Choose Your Guide:

  • VSCodium Web IDE: The absolute best way to code in the cloud. Spin up a browser-based editor that is completely isolated and secured by SSO.
  • OpenClaw: Give the powerful autonomous AI agent a safe, disposable sandbox to execute code in.
  • Hermes: Run agent-driven background services and expose only the user interface to the web.
  • Docker Compose Stacks: The cleanest way to run multi-container apps (like a web server + Postgres + Redis) without making a mess of your host machine.
  • Dokploy: Turn your Terrarium containers into an automated, UI-driven deployment platform (like Heroku or Vercel).
  • Coolify: Another excellent, self-hosted Heroku alternative that runs perfectly inside a single Terrarium LXC.
  • Protecting Services with OIDC: The magic label that forces users to log in before they can see your published web apps.

Built with VitePress