Your First Session

This walkthrough goes from a fresh install to a running agent in under a minute. It assumes you’ve installed Verne and have at least one CLI agent (such as claude) on your PATH.

1. Open a folder

The welcome screen greets you on first launch. Press ⌘O (or click Open Folder) and choose a project directory, usually the root of a git repository.

The folder becomes a workspace and stays there across restarts. Open as many workspaces as you like; they’re listed in the Workspaces panel on the left. See Workspaces & worktrees for the full model.

2. Launch your agent

Opening a folder creates a terminal tab rooted in that directory. The terminal is a real shell, so run anything you normally would.

To start an agent, run its command:

$ claude

Verne recognizes the agent and starts tracking it. For supported agents the integration is set up for you, so there’s nothing to configure. See Agents & status.

3. Watch the status

Once an agent is running, a status icon appears on its tab and in the Agents panel on the left:

  • Working: a spinning blue loader
  • Blocked: an amber alert icon, waiting for your input
  • Done: a purple check
  • Idle: a green check

The Agents panel lists every agent across all your workspaces, so you can keep an eye on several at once. Click any entry to jump to its tab. When an agent finishes in the background, Verne flags its tab so you don’t miss it.

4. Bring in your files

Verne gives the agent a whole workspace to work in. While it runs you can:

  • Browse and edit code in the editor (Monaco, with LSP).
  • Review changes and commit in Source Control.
  • Hand a file to the agent: drag it from the file explorer onto the terminal, or press ⇧⌘L in the editor to inject the current file path into the agent’s input. See Add files to an agent.

Next steps