First-Run Welcome

What happens when you run getbot for the first time.

demo

getbot access is invite-only. The CLI and docs are public. Deployment via getbot setup requires an invite code. Request an invite code if you don't have one.

Your first encounter

Run getbot with no arguments on a fresh machine and you'll get a welcome message that explains what getbot does and what you need to get started. It's not a help dump — it's a warm handshake.

$ getbot

  Welcome to getbot!

  Deploy your own AI agent — secured in an Incus container,
  accessible via HTTPS, authenticated with Google SSO.

  You'll need:
    • A server running Ubuntu 22.04+, Debian 12+ with SSH access (or a DigitalOcean account)
    • An API key from Claude, ChatGPT, or Gemini

  Ready to get started? [Y/n]:

Press Enter (or type y) and getbot chains directly into the setup wizard. If you're not ready, type n and come back later with getbot setup.

Non-interactive environments

In a CI pipeline or when stdin isn't a terminal, getbot skips the interactive prompt and prints a one-liner instead:

Run 'getbot setup' to get started.

For non-interactive deployments, run getbot setup --yes with all required flags. See the CI Automation docs (coming in a later batch) for full details.

What you need

A Linux server — any VPS or bare metal with SSH access. Ubuntu 22.04/24.04 or Debian 12/13. getbot handles all the setup from there.
An LLM API key — from Anthropic (Claude), OpenAI (ChatGPT), or Google (Gemini).
A work email address — getbot derives your organization name from the email domain. Personal email providers (gmail, outlook, etc.) require the --org flag.

If you have a DigitalOcean account, getbot can auto-provision a droplet for you. But SSH-first is the primary path — bring your own server and getbot does the rest.