Privacy-First AI
The Mac File Organizer That Never Touches Your Data
Most AI tools send your files to a remote server to analyse them. OCDify runs its entire AI pipeline on your Mac — so your documents, source code, and private files stay exactly where they are.
Your files never leave your Mac
OCDify reads file names and contents to classify them. That analysis happens exclusively on your device using a locally running AI model. No file content, metadata, or path is sent to any server — ours or anyone else's.
Offline AI — no internet connection needed
OCDify downloads its AI model (gemma-3 via llama.cpp, ~1 GB) once on first launch. After that, the app runs fully offline. You can block its network access entirely in macOS's firewall and every feature continues to work.
No account, no sign-in
There is nothing to register for. Download the app, activate your trial, and start organising. No email address, no password, no profile — there is no server storing your identity.
License verified locally
Your license key is cryptographically signed using ed25519. The app verifies the signature against a public key embedded in the binary — with no network check-in. You can activate your license offline.
No analytics or tracking in the app
The OCDify desktop app collects zero usage data. There are no analytics SDKs, no crash reporters of our own, and no telemetry. Apple's system crash reporting applies only if you've opted into it in System Settings.
Runs in a macOS sandbox
OCDify is a sandboxed macOS app and runs as a background agent (LSUIElement). It requests file access only through macOS's standard permission prompts. You can revoke access to any folder at any time.
How the on-device AI actually works
Model download (once)
On first launch OCDify downloads the gemma-3-1b model from Hugging Face (~1 GB). This is the only network request the AI component ever makes. After this download you can use OCDify offline indefinitely.
Local inference via llama.cpp
OCDify uses llama.cpp to run the model natively on your Mac's CPU or Apple Silicon. The model reads the file's name and a sample of its content, then outputs a category label. This runs in memory on your machine.
File moved, nothing logged
Once the AI picks a category, OCDify moves the file to the matching subfolder. No record of the file name, path, or category is sent anywhere. The operation is purely local.
Model cached locally
The model stays in ~/Library/Containers/com.vyanlabs.ocdify. It is never re-downloaded unless you reinstall. Deleting the app and its container removes everything OCDify stored on your machine.
OCDify vs. cloud-based AI file organizers
| Feature | OCDify | Cloud AI tools |
|---|---|---|
| AI runs on-device | ✓ | ✗ |
| Files uploaded to server for analysis | ✗ | ✓ |
| Account / sign-in required | ✗ | ✓ |
| Works fully offline | ✓ | ✗ |
| Usage analytics collected from you | ✗ | ✓ |
| Monthly subscription | ✗ | ✓ |
| License validated via internet | ✗ | ✓ |
Who benefits most from offline AI file organization
Freelancers & consultants
Client contracts, NDAs, and project files shouldn't pass through a third-party server. OCDify keeps them organised without leaving your machine.
Developers
Source code, API keys in config files, and private repos deserve to stay local. OCDify can sort your dev downloads without ever reading them remotely.
Finance & legal professionals
Regulated industries with data handling obligations benefit from a tool with zero data transmission — no BAAs, no compliance headaches.
Privacy-conscious individuals
If you don't want your personal documents analysed by a tech company's servers, OCDify is the only AI file organizer that makes that possible.
Private by design. Automatic by nature.
Try OCDify free for 30 days. No account, no credit card, no cloud. Best installed via the Mac App Store.
Recommended · Auto-updates · Family Sharing
Read our full privacy policy for every technical detail.