wt

wt

A fast CLI tool for managing git worktrees with fuzzy selection.

Features

• Quick worktree creation with optional branch name preprocessing
• Interactive fuzzy finder for navigating between worktrees
• Multi-select deletion of worktrees
• File copying with gitignore-style patterns
• Post-creation hooks for automated setup
• Shell integration for seamless directory switching
• Tmux support for opening worktrees in new panes

Installation

Homebrew

brew install default-anton/tap/wt

From releases

Download the latest release from GitHub Releases.

From source

go install github.com/default-anton/wt/cmd/wt@latest

Shell Setup

For wt cd and wt add to automatically change your directory, add shell integration.

Note: If installed via Homebrew, shell integration and completions are set up automatically. You can skip this section.

Bash

# Add to ~/.bashrc
eval "$(wt shell-init bash)"
eval "$(wt completion bash)"

Zsh

# Add to ~/.zshrc
eval "$(wt shell-init zsh)"
eval "$(wt completion zsh)"

Fish

# Add to ~/.config/fish/config.fish
wt shell-init fish | source
wt completion fish | source

Usage

Create a worktree

# Create worktree with branch name
wt add my-feature

# With tmux (opens in new pane)
wt add my-feature -t  # or --tmux

# With custom base branch
wt add my-feature --base develop

Go to a worktree

# Interactive fuzzy finder
wt cd

# With tmux
wt cd -t  # or --tmux

Remove worktrees

# Interactive multi-select
wt rm

# Direct removal
wt rm .worktrees/my-feature

# Force removal
wt rm -f .worktrees/my-feature

List worktrees

wt ls

Initialize config

wt init

Configuration

Run wt init to create a .wt.toml configuration file in your repository root. This command also adds the worktree directory to .gitignore.

Example configuration:

# Base branch for new worktrees (default: main)
base_branch = "main"

# Directory for worktrees (default: .worktrees)
worktree_dir = ".worktrees"

# Preprocessing script (receives input, outputs branch name)
preprocess_script = ".wt/preprocess.sh"

# Files/directories to copy (gitignore-like patterns)
# Supports ** for recursive matching (e.g., **/node_modules for monorepos)
copy_patterns = [
  "**/node_modules",
  ".env*",
  "vendor",
  "!.env.example",
]

# Post-creation hooks
[[post_hooks]]
name = "Install dependencies"
run = "npm install"

[[post_hooks]]
name = "Setup database"
run = "bin/rails db:prepare"
if_exists = "bin/rails"

Preprocessing Script

You can define a script that transforms the input into a branch name. This is useful for extracting branch names from issue tracker URLs:

#!/bin/bash
# .wt/preprocess.sh

# Example: Extract JIRA ticket from URL
# Input: https://jira.example.com/browse/PROJ-123
# Output: PROJ-123-feature-name

INPUT="$1"

# Extract ticket ID from URL or use as-is
if [[ "$INPUT" =~ ([A-Z]+-[0-9]+) ]]; then
  TICKET="${BASH_REMATCH[1]}"
  echo "$TICKET"
else
  echo "$INPUT"
fi

Make sure the script is executable: chmod +x .wt/preprocess.sh

License

MIT License - see LICENSE for details.