> ## Documentation Index
> Fetch the complete documentation index at: https://docs.cubby.pro/llms.txt
> Use this file to discover all available pages before exploring further.

# Choose your path

> Pick the fastest Cubby workflow for your app and audience

# Choose your path

Cubby is intentionally small: install the CLI, connect an app, deploy it, and share the URL. The best first steps depend on what you are trying to ship.

<CardGroup cols={2}>
  <Card title="AI coders" icon="wand-magic-sparkles" href="/tutorials/expert-from-paste">
    Use Codex, Claude, Cursor, or another agent. Paste one prompt and it installs the CLI, runs `cubby init`, reads the generated rules, and deploys.
  </Card>

  <Card title="Rapid prototypes" icon="bolt" href="/tutorials/first-app">
    Start from `cubby init`, deploy early, and iterate in public or with a small invite-only group.
  </Card>

  <Card title="Database-backed apps" icon="database" href="/tutorials/database-apps">
    The default template is **SQLite**; choose **Neon Postgres** when you need Postgres. Both come with Prisma wired up.
  </Card>

  <Card title="Shared tools" icon="users" href="/platform/sharing">
    Invite collaborators, manage access, and let Cubby handle authentication around the app.
  </Card>
</CardGroup>

## If you are using an AI coding assistant

You do not need a working local app first — let the agent scaffold one. Paste the
[Expert from paste](/tutorials/expert-from-paste) prompt into Codex, Claude, or Cursor
and it will install the CLI, log in, run `cubby init` with the right template, read the
generated `CUBBY.md` / `AGENTS.md` / `CLAUDE.md`, ask you about local dev, then
`cubby check` and `cubby deploy`.

Per-agent detail:

1. [Expert from paste](/tutorials/expert-from-paste) — the agent-neutral prompt
2. [Using Cubby with Codex](/tutorials/using-with-codex)
3. [Using Cubby with Claude](/tutorials/using-with-claude) — adds an optional MCP path
4. [Using Cubby with Cursor](/tutorials/using-with-cursor)

## If you are bringing an existing app

Use Cubby when your app can run as a web process and expose a health endpoint. Before deploying, confirm:

* The app has a deterministic build command.
* The app listens on the port Cubby provides.
* Required environment variables are documented.
* Secrets are configured through Cubby, not committed to git.
* The app has a route you can use to confirm it is live after deploy.

See [platform concepts](/platform/concepts) for the mental model and [Cubby configuration](/platform/cubby-md) for project metadata.

## After your first deploy

* Share the app with teammates or friends: [Sharing](/platform/sharing)
* Make it feel native on phones and desktops: [PWA support](/platform/pwa)
* Understand authentication boundaries: [Authentication](/platform/authentication)
* Estimate plan limits and costs: [Billing](/platform/billing)