# autojump — Shell Directory Navigation That Learns Your Habits

> A Python-based tool that maintains a database of your most-visited directories and lets you jump to them with a short partial name.

## Install

Save as a script file and run:

# autojump — Shell Directory Navigation That Learns Your Habits

## Quick Use
```bash
# Install
brew install autojump        # macOS
sudo apt install autojump    # Debian/Ubuntu

# Add to shell config (~/.bashrc or ~/.zshrc)
# source /usr/share/autojump/autojump.sh

# Jump to a frequently used directory
j project    # jumps to /home/user/code/my-project
```

## Introduction
autojump is a faster way to navigate your filesystem from the command line. It works by maintaining a weighted database of directories you visit through cd, then allowing you to jump to the best match using just a few characters of the directory name.

## What autojump Does
- Tracks directory visit frequency in a local database updated on every cd command
- Provides the ``j`` command for jumping to the highest-weighted matching directory
- Supports partial name matching and multiple arguments for disambiguation
- Opens directories in your file manager with the ``jo`` command
- Works with bash, zsh, and fish shells

## Architecture Overview
autojump is written in Python and hooks into your shell via a function that wraps cd. Each time you change directories, the wrapper increments the weight for that path in a local text-based database stored in ~/.local/share/autojump/. When you invoke ``j``, the tool scans the database for the best match by weight and recency, then changes to that directory.

## Self-Hosting & Configuration
- Install via package managers (Homebrew, apt, pacman) or from source
- Source the autojump shell script in your shell's rc file to enable the cd hook
- The database file is plain text and can be manually edited or backed up
- Configure ignored directories by adding paths to ~/.config/autojump/
- Works without root privileges; all data lives in the user's home directory

## Key Features
- Zero-configuration learning: starts working immediately after installation
- Handles partial directory names and fuzzy matches
- Supports tab completion for multiple match candidates
- Lightweight with minimal shell startup overhead
- Cross-platform: runs on Linux, macOS, and Windows (via Cygwin or WSL)

## Comparison with Similar Tools
- **zoxide** — Rust rewrite inspired by autojump; faster startup and more active development
- **fasd** — tracks files and directories but is no longer actively maintained
- **z (rupa/z)** — similar concept in pure shell script; lighter but fewer features
- **fzf** — general fuzzy finder that can be combined with cd for directory jumping but requires more setup
- **broot** — interactive directory navigator with tree view; different workflow paradigm

## FAQ
**Q: Does autojump slow down my shell startup?**
A: The overhead is minimal. Sourcing the script adds a few milliseconds to shell initialization.

**Q: What happens if I rename or delete a directory?**
A: autojump will try to match the old path and fail silently. You can purge stale entries with ``j --purge``.

**Q: Can I use it alongside zoxide?**
A: They serve the same purpose and would conflict on the ``j`` alias. Pick one or configure different aliases.

**Q: Does it work in fish shell?**
A: Yes. autojump includes a fish plugin that integrates with the fish completion system.

## Sources
- https://github.com/wting/autojump
- https://github.com/wting/autojump/wiki

---
Source: https://tokrepo.com/en/workflows/asset-3d7fd8c9
Author: Script Depot