--- title: Private Mac File Organizer with Rules & AI | Mochi url: https://www.trymochi.dev/ --- A private file organizer for Mac # Organize files. Without the busywork. Mochi quietly tidies Downloads, Desktop, and the folders you choose using clear local rules you can review and undo. Optional AI can suggest where ambiguous files belong, but it never moves a file without your approval. [Download for Mac](https://github.com/jerrysun-ofive/mochi/releases/latest/download/Mochi.dmg) ## A rule when you need it. Choose only the folders Mochi may access, start with useful sorting rules, preview every proposed move, and undo later if you change your mind. ## Private by design. Core organization happens on your Mac. AI is off by default, your keys stay in macOS Keychain, and a local Ollama model can keep optional classification offline. ## Free and open source. Mochi supports macOS 13 or later on Apple silicon and Intel Macs. Inspect the source, release notes, privacy architecture, and SHA-256 checksum before installing. ## Choose folders, not your whole Mac. Add Downloads, Desktop, or another location individually. Mochi stores macOS security-scoped access only for folders you authorize. Removing a watched folder stops new organization without deleting the folder or its files. Rules can match extensions, names, dates, Finder tags, and other metadata. Start with a narrow rule such as moving DMG and PKG files into Installers. Run it in review mode, inspect why each file matched, and refine the condition before expanding the scope. ## Uncertainty stays visible. Potential moves can wait in a review queue. You see the proposed destination and reason before approval. Completed operations appear in local history, and eligible moves can be undone when the source and destination remain safe. Mochi never silently overwrites an existing file. Optional AI runs only after ordinary local rules cannot decide. It chooses from destinations supplied by the Mac client and cannot browse the filesystem or execute a move. Use supported provider credentials stored in macOS Keychain, or keep classification local with Ollama. ## Verify the current release. Version 0.1.1 is approximately 4.1 MB and supports Apple silicon and Intel Macs. The canonical GitHub release publishes the DMG, release notes, manifest, and SHA-256 checksum. Compare that checksum before installing and confirm that macOS identifies the expected developer. Mochi is built by Jerry Sun as an open-source project. The public repository includes the AI architecture, benchmark plan, security policy, distribution contract, source history, and issue tracker. Support is available at trymochi.dev@gmail.com. ## Privacy claims you can inspect. Core organization happens locally. Rules, review decisions, undo history, and folder bookmarks stay on the Mac. AI is off until you configure it. When an external provider is enabled, Mochi prepares a limited description only after local rules cannot decide, redacts common personal patterns on the Mac, and sends the request directly to the provider you selected. The model returns a suggestion rather than a filesystem command. The Mac client checks the response, limits it to your known destinations, and keeps the move behind review. Complete files, absolute paths, provider credentials, and the local operation journal remain outside that provider boundary. A local Ollama model can keep optional classification on the Mac. ## Frequently asked questions ### Is Mochi free? Yes. The Mac app is free and open source under the Apache 2.0 license. Rules, watched folders, review, history, undo, and bring-your-own-key AI are all included. Mochi Pro, a hosted AI option planned at A$8 per month, is not available yet and is not required for anything. ### Does Mochi upload my files? No. Rules run on your Mac, and file contents, filenames, and folder paths are never sent to Mochi’s servers. AI is off by default. If you turn it on, Mochi sends a limited, locally redacted description of a file that rules could not place directly to the AI providers you configured. Complete files are never uploaded. ### Does Mochi work offline? Yes for rules. Watched folders, rules, review, history, and undo work offline and need no account. Optional AI suggestions need a network connection to the providers you configure. ### Will Mochi move files without asking? Proposed moves go to a review queue where you see the file, the destination, and the reason before approving. AI can only suggest a destination from folders you chose; it can never approve or perform a move by itself. ### Can I undo a move? Yes. Every completed move is recorded in local history, and eligible moves can be undone. Mochi never silently overwrites an existing file; if undoing would collide with another file, it stops instead. ### Which Macs does Mochi support? macOS 13 Ventura or later, on both Apple silicon and Intel Macs. The download is approximately 4.1 MB. ### How is Mochi different from Hazel? Hazel is a mature, paid automation app with powerful scripting. Mochi is free and open source, keeps proposed moves in a review queue with undo history, and can optionally ask AI about files that rules cannot place. [Read the full Mochi vs Hazel comparison](https://www.trymochi.dev/hazel-alternative) ### Is Mochi ready for important files? Mochi is in early development (version 0.1.1). Keep backups, start with a low-risk folder such as Downloads, and review moves before approving them while you build confidence in your rules. ## Guides - [Mochi vs Hazel](https://www.trymochi.dev/hazel-alternative) - [Organize Downloads](https://www.trymochi.dev/organize-downloads-mac) - [Organize screenshots](https://www.trymochi.dev/organize-screenshots-mac) - [Tidy your Desktop](https://www.trymochi.dev/organize-desktop-mac)