jhadoo

Auto-clean unused development environments, installers, cache files, and remnants. Built to run natively across macOS, Windows, and Linux.

AI-assisted coding tools (like Cursor, Copilot, Bolt, or Windsurf) make prototyping incredibly fast, but they leave behind a massive footprint of heavyweight folders (venv, node_modules, build caches) for every small experiment. A few weeks of coding can silently eat 20 to 50 GB of disk space.

Jhadoo is a lightweight, cross-platform terminal utility that scans and purges stale workspace environments, package caches, local installers, and leftover app configurations. It helps maintain a fast, continuous vibe coding workflow without manual filesystem housekeeping.

Quick Start (Select Your OS)

macOS & Linux

pipx install jhadoo && jhadoo --dry-run

Windows (PowerShell)

pipx install jhadoo; jhadoo --dry-run

Supported Operating Systems

Jhadoo provides native, equal-parity integration across:

• macOS: Scans app bundles, cleans CocoaPods/Homebrew caches, and supports macOS system paths.
• Windows: Queries registry uninstall entries, purges %TEMP%, and includes fallback support for Windows console contexts.
• Linux: Integrates with package managers (apt, dnf, pacman) and cleans up user configurations (~/.config, ~/.local).

Core Features

• Terminal Dashboard (jhadoo --tui) - A fully interactive 24-bit Truecolor console interface:
  • Disk Tree Explorer: Drill down into directories sorted by size and flag files or folders for archiving using standard Arrow Keys or Vim hotkeys (h/j/k/l).
  • Live Telemetry: Monitor logical CPU utilization, memory metrics, and I/O rates alongside a calculated system health index.
• System-Wide Optimizer (jhadoo --optimize) - Flushes local DNS caches, purges system temporary directories, and updates standard workspace configurations.
• Application Uninstaller (jhadoo --uninstall [APP_NAME]) - Removes designated applications along with deep configurations, plists, launch daemons, and hidden leftovers.
• Installer Sweeper (jhadoo --installers) - Identifies and safely purges leftover .dmg, .pkg, .msi, .exe, .deb, and .rpm files in Downloads and Desktop directories.
• Smart Staleness Verification - Evaluates true project activity. Skips folders modified solely by automated OS metadata (like .DS_Store or Thumbs.db) that trigger false freshness values.

Zero-Risk Safety Guardrails

Jhadoo is built with conservative defaults to protect system directories and critical user data:

1. System Path Guardian: Scans are explicitly restricted from modifying critical OS-specific directories (such as /System on macOS, C:\Windows on Windows, or /boot and /etc on Linux).
2. Dry-Run Mode (--dry-run / -n): Preview exactly what folders, caches, or applications will be targeted before any changes occur on disk.
3. Archive and Restore (--archive / --restore): Instead of permanently deleting assets, Jhadoo can move them to a secure local archive directory (~/.jhadoo_archive). Use jhadoo --restore to instantly revert the previous session's cleanup.
4. No Greedy Wildcards: Leftover config directories are identified strictly using bounded package or vendor matches, keeping shared directories safe.

Installation

Standard Installation

# Recommended (install as a global, isolated binary)
pipx install jhadoo

# Standard pip installation
pip install jhadoo

Note: For Windows users wanting to run the interactive curses TUI, please install the dependency wrapper:

pip install windows-curses

Run Directly from the Source Code (GitHub Repo)

If you want to run, contribute, or test the tool directly from the git repository on your system without installing the package globally, follow this procedure:

1. Clone the Repository:

   git clone https://github.com/bhav09/jhadoo.git
   cd jhadoo

2. Set Up a Virtual Environment (Optional but recommended):

   python3 -m venv .venv
   source .venv/bin/activate  # On Windows, use: .venv\Scripts\activate

3. Install in Editable Mode: This registers the jhadoo command locally and links it directly to your source files:

   pip install -e .

4. Alternatively, Run as a Python Module: If you do not want to install it at all, you can invoke the CLI entry module directly:

   python3 -m jhadoo --dry-run

Command Reference

Default Scanning and Core Cleaning

jhadoo                # Run the standard developer project cleanup scan
jhadoo --dry-run      # Safely preview potential savings without modifying anything
jhadoo --archive      # Back up matching targets to ~/.jhadoo_archive instead of deleting
jhadoo --restore      # Revert the last run's archived deletions back to original paths
jhadoo --dashboard    # View past savings history, clean statistics, and trend predictions

Specialized Operations

jhadoo --tui         # Launch the interactive console disk and telemetry dashboard
jhadoo --optimize    # Run system maintenance (flush DNS, purge temp files, clean package stores)
jhadoo --installers  # Scan Downloads and Desktop for setups and installer binaries
jhadoo --uninstall   # Interactive prompt to remove applications and their configuration remnants

Privacy and Telemetry

• Local Processing: File paths, usernames, and system details never leave your computer.
• Minimal Telemetry: To track the overall open-source impact, Jhadoo runs aggregate telemetry metrics (global bytes freed, runtime, and OS type) securely enabled on installation.
• Opt-Out Anytime:
  • Disable: jhadoo --telemetry-off
  • Check status: jhadoo --telemetry-status

QA and Troubleshooting

macOS DNS flush (non-admin)

Running jhadoo --optimize without sudo may report a partial DNS flush: dscacheutil succeeds, but reloading mDNSResponder requires administrator privileges. This is expected. For a full flush, run sudo jhadoo --optimize.

Manual Ctrl+C testing

On fast storage, cleanup can finish before you can press Ctrl+C. Use a deliberate delay:

jhadoo --archive --test-delay 10 --config /path/to/test-config.json

Point the config main_folder at a small stale test project. Expect exit code 130 and a partial deletion manifest.

PyPI old versions

Jhadoo follows a rolling retention policy: only the latest 3 PyPI releases are kept installable by default. Older versions are yanked on a best-effort basis.

Verify the current yank state:

python scripts/yank_old_pypi_versions.py --verify-only --keep 3

Why explicit pip install jhadoo==1.2.0 may still succeed (and why that's not a Jhadoo bug):

Per PEP 592, "yanked" releases on PyPI are only hidden from default resolution. An explicit pin (==1.2.0) will still install a yanked file — pip prints a warning but proceeds. The only way to fully block installation of an old version is to delete the release on PyPI within 72 hours of upload, after which deletion is permanently disallowed by PyPI policy. There is no programmatic API to yank or delete releases on PyPI; maintainers must do this manually via the PyPI web UI at https://pypi.org/manage/project/jhadoo/releases/.

The expected behaviour that is enforced:

• pip install jhadoo → resolves to the latest non-yanked release (always within the last 3).
• pip install -U jhadoo → never crosses a yanked version boundary.

If --verify-only reports older versions as still active, the maintainer has not yet yanked them in the PyPI UI — please open an issue at https://github.com/bhav09/jhadoo/issues so we can complete the manual yank.

License

MIT License. See LICENSE for more details.