Quick start

rove runs many AI coding sessions side by side in your terminal. Each managed Task gets its own git worktree and branch, so parallel Tasks never step on each other. Extra tabs inside one Task share that Task's directory.

You need git and at least one engine CLI (claude, codex, copilot, or kimi) on your PATH. The Rove CLI itself runs on the Bun runtime (≥ 1.3.11) — you do not have to install Bun yourself, every route below brings it along.

Windows also requires Node.js and Git for Windows. Rove uses Node.js for its Windows PTY host and Git Bash as the POSIX shell behind engine and terminal tabs. Restart Rove after installing either one so the new executables are on PATH.

Install

One line on a fresh machine — installs Bun if it is missing, then Rove:

curl -fsSL https://rove.sma1lboy.me/install.sh | sh

# pin a version
curl -fsSL https://rove.sma1lboy.me/install.sh | sh -s -- 0.8.136

Or use a package manager you already have:

npm install -g @sma1lboy/rove   # npm — the first launch offers to install Bun
bun install -g @sma1lboy/rove   # bun
npx @sma1lboy/rove              # try it without installing

On Windows, use the npm route (the shell script needs a POSIX shell):

npm install -g @sma1lboy/rove

Rove looks for Bun on PATH, in $BUN_INSTALL/bin, and in ~/.bun/bin. If yours lives somewhere else, point Rove at it with ROVE_BUN=/path/to/bun. To never be asked about installing Bun (CI, images, locked-down machines), set ROVE_NO_BUN_BOOTSTRAP=1 — Rove then prints the install commands and exits.

Install the agent skill

Do this in the same sitting as the install above. The companion skill teaches your coding agent (Claude Code, Codex, …) how to drive Rove itself — spawn tasks, fan a prompt out to several attempts, compare them, land the winner:

rove skill install

Claude Code users have a one-stop alternative: the Rove plugin carries the skill AND the activity hooks in one install, with no PATH or settings.json setup:

/plugin marketplace add Sma1lboy/rove
/plugin install rove@rove

If you were already running Rove before installing the plugin, run rove hook cleanup once afterwards — details in Configuration → Claude Code plugin.

First launch

The first rove launch asks two optional setup questions: whether to install shell completions for the detected shell, and whether to install the companion Rove skill for coding agents. Choose with j/k or the arrow keys and confirm with enter. q or esc skips anything you have not answered.

The wizard finishes with a short keyboard primer. It does not sign in to an engine; install and authenticate at least one supported engine CLI separately. The skill question is the same rove skill install from the section above — answering it there is enough.

Your first task

cd your-repo
rove

Press n, pick a repo, a base branch, and an engine. Then just talk to the session — it's the real engine CLI, running in a fresh worktree under ~/.rove/worktrees/ (existing ~/.kobe/worktrees/ tasks remain usable).

Three panes: tasks on the left, the engine session in the middle, changed files on the right. Click any of them to focus it.

Three keys to remember

Quitting doesn't stop anything

Sessions keep running in the background after you quit, close the terminal, or drop an SSH connection. Run rove again and everything is where you left it. Finishes, failures, and approval requests are recorded in the Inbox and show as unread when you return.

Desktop notifications need an attached Rove TUI and a terminal that supports OSC 9. They can ride an active SSH connection to your local terminal, but once the terminal or SSH stream is gone there is nowhere to send one; the durable Inbox entry is the notification you see on the next attach.

Run many attempts at once

One prompt, N isolated attempts, one command:

rove api add --repo "$PWD" \
  --agents claude:2,codex:2 \
  --prompt "Try independent approaches to simplify the auth flow."

Compare the attempts, then land the winner:

rove api collect --task-ids a,b,c      # read-only comparison
rove api land --task-id a              # merge the winning branch

Let your agent drive

With the agent skill in place, you do not have to type those commands yourself — ask your coding agent for three attempts at a prompt and it runs the rove api loop above for you, then reports back which branch to land.

If something's wrong

rove doctor            # check daemon, engines, git
rove doctor --report   # write a bundle for a bug report

More fixes in Troubleshooting.

Next steps

• Concepts — tasks, sessions, and what survives what.
• The TUI — status glyphs, the Inbox, diff review, and the pages.
• CLI reference — every rove command.
• rove api — the scriptable surface for scripts and agents.
• Configuration — engines, themes, notifications.