🧹 Junk Cleaner

An elegant, AI-powered system cleaner built entirely on local Large Language Models (LLMs). No cloud needed.

✨ Features

• AI-Generated System Reports: Get a readable health summary of your entire system volume directly from a local LLM.
• Model Deduplication: Find and isolate heavy .gguf / .safetensors weights duplicated across Ollama and LM Studio — using SHA-256 checksums.
• Smart App Uninstaller: Not just .app deletions — parses .plist files and deep-scans Application Support and Preferences for lingering data.
• Developer Cleanup: Discovers regeneratable build artefacts (node_modules, target, .venv, .next, __pycache__, DerivedData, and 100+ more) across 15 ecosystems. Shows the exact command to regenerate each folder so you can delete safely.
• Copilot Chat: Type natural language commands ("scan for junk", "find large files") — or just chat freely with the local AI. Non-scan queries get real conversational replies.
• Glassmorphism UI: Polished dark interface with animated micro-interactions, a custom model picker with ⭐ Recommended badges, and consistent component sizing.
• Strictly Read-Only & Secure: Multiple layers of write-protection — Tauri capability whitelist, Rust-side action whitelist, frontend whitelist, hardened AI system prompts, and localhost-only Ollama enforcement.

🚀 Getting Started

To get started quickly, download the latest installer from our Releases page. We build native installers for macOS, Windows, and Linux.

Note: Clicking the buttons above will take you to our latest GitHub release. Download the appropriate .dmg (Mac), .exe (Windows), or .AppImage (Linux) file for your machine.

⚡ Automated CLI Install

For a lightning-fast install that automatically bypasses warnings, open your terminal (or PowerShell) and paste the command for your OS:

macOS & Linux (Terminal):

curl -fsSL https://raw.githubusercontent.com/Divish1032/junk-cleaner/main/install.sh | bash

Windows (PowerShell):

irm https://raw.githubusercontent.com/Divish1032/junk-cleaner/main/install.ps1 | iex

🍏 macOS Installation Note

Because this app is currently open-source and not yet digitally signed with a paid Apple Developer account, macOS Gatekeeper may show a warning saying "Apple could not verify JunkCleaner is free of malware".

To open it:

1. Open your Applications folder in Finder.
2. Right-click (or Control-click) the JunkCleaner app icon.
3. Select Open from the context menu.
4. Click Open again on the pop-up warning. (You only need to do this once!)

Local AI Features

Junk Cleaner utilizes your local processing power for its AI features (Chat, Smart Classification, Health Reports). The application includes a built-in onboarding experience that will guide you through installing and connecting to Ollama seamlessly if you don't already have it!

🛠 Building from Source

If you want to contribute or build the application from source, make sure you have Node.js 18+ and the Rust tooling installed.

# Clone the repository
git clone https://github.com/<your-username>/junk-cleaner.git

# Navigate into the project
cd junk-cleaner

# Install Frontend dependencies
npm install

# Run the Tauri Developer Environment
npm run dev

To build a production bundle manually:

npm run tauri build

🤝 Contributing

We welcome all contributions! Whether it's porting to a new platform, fixing a UI bug, or writing a better Rust implementation for the backend - we'd love to have you.

To get started, please check out our Contributing Guide and adhere to our Code of Conduct.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.