DiskMap

See where your disk went, and what you haven't touched in months.
A fast, private disk map for macOS with an AI you can ask "is this safe to delete?"

_{Screenshots use a demo disk with made-up files.}

What it does

Finds what's big and forgotten. Every folder is a colored tile sized by how much space it takes. Tiles with diagonal hatching haven't been opened for N days (you set N with the slider). The headline tells you the total up front.

Reads your whole Mac in about a minute. A parallel Rust scanner read 9 million files in about 60 seconds on an Apple Silicon Mac, with a live progress bar measured against your disk's real usage. The app opens on your last scan, so you only rescan when you want fresh numbers.

Searches the whole scan. Type keywords to match paths across every folder. You get the total size of the matches, and the map dims everything else.

Lets you ask an AI. Chat about your disk with a local model through Ollama, or with OpenAI. It sees a summary of your scan, so you can ask "why is my disk so full?" or "what can I safely delete?" It can propose deletions, but it can never delete anything on its own.

Deletes safely. Nothing is removed until you confirm in a dialog that lists exact paths, sizes and how long since you last used them. Items go to the Trash, so you can put them back. System folders, your home folder, and credentials are blocked.

Also: a Since last scan tab showing what grew or shrank, category chips (node_modules, Python envs, Docker, caches, AI models, Xcode), and light, dark, or automatic appearance.

Install

1. Download DiskMap_<version>_*.dmg from the Releases page.
2. Open it and drag DiskMap into Applications.
3. The app isn't signed with an Apple Developer ID yet, so macOS blocks the first launch. Either right-click the app and choose Open, or run:

   xattr -dr com.apple.quarantine /Applications/DiskMap.app

4. For a complete count, give it Full Disk Access (System Settings > Privacy & Security > Full Disk Access). Without it, protected folders are skipped and the totals run low. You only need to do this once.

Requires macOS 11 or later. The current download is for Apple Silicon Macs; the release workflow is set up to build a universal version that also runs on Intel.

AI chat

Click the sparkle button in the toolbar and choose a provider.

	Ollama	OpenAI
Where it runs	On your Mac	OpenAI's servers
Setup	Start Ollama, pick a model	Paste an API key, pick a model
Key storage	none needed	macOS Keychain, never a file

What is sent: the folder path (your home folder shown as ~), sizes, dates, and the names of files inside it. Never file contents. With Ollama, nothing leaves your Mac.

Build from source

You need Node and Rust.

npm install
npx tauri build --bundles dmg   # installer in src-tauri/target/release/bundle/dmg
npx tauri dev                   # run in development
cd src-tauri && cargo test      # tests

Releasing

Push a version tag and GitHub Actions builds a universal (Apple Silicon + Intel) DMG and attaches it to a draft release:

git tag v0.1.0 && git push origin v0.1.0

How it works

A Rust backend (Tauri) walks the disk with a parallel directory scanner and returns a size-pruned tree; items under 10 MB are folded together. The interface is plain HTML, CSS and JavaScript, with d3-hierarchy for the treemap. There is no web server and no account: the scan stays on your machine.