TL;DR β Hazel vs FilesDesk in 30 seconds
- Hazel is a mature rules engine for macOS. You write conditions, it executes actions. Predictable, scriptable, Mac-only.
- FilesDesk is an AI file organizer for Mac and Windows. It reads what's inside each file and renames it β no rules to write or maintain.
- Pick Hazel when your files arrive in predictable formats and you want deterministic control on a Mac.
- Pick FilesDesk when your files are too messy or varied to describe in rules, or when you also need Windows.
- Many power users run both: Hazel routes, FilesDesk names.
If you've ever Googled "how do I get Hazel for Windows," or watched a perfectly tuned Hazel rule quietly stop working the day a vendor changed their invoice template, this post is for you.
Hazel has been the gold standard for Mac file automation for over fifteen years. It's a rules engine β it does exactly what you tell it, every time, without interpretation. FilesDesk takes a different approach entirely: hand the file to an AI, let it read what's inside, and produce a clean name based on the actual content. No rules. No regex. No maintenance when a layout changes.
This is a fair, side-by-side look at what each tool actually does, where each one earns its keep, and how to decide β or how to use them both without stepping on each other's toes.
What Is Hazel?
Hazel is a macOS automation app from Noodlesoft. It watches folders you designate β Downloads, Desktop, a scanner inbox, anywhere β and runs rules against every file that lands there. A rule has two parts: conditions (if the file name contains "invoice" and the kind is PDFβ¦) and actions (rename it, move it, tag it, run a shell script, trash it).
That's the whole model, and it's genuinely powerful when your files behave. Where Hazel shines:
- Deeply mature and stable β over fifteen years of refinement and a devoted following
- Deterministic: the same file, the same conditions, the same result, every time
- Extremely capable when file patterns are predictable and consistent
- First-class AppleScript and shell script integration for anything it can't handle natively
And where it hits its limits:
- Mac-only. There is no Windows version, and Noodlesoft hasn't indicated one is coming.
- Every rule has to be written by a human and maintained by a human. When a sender changes their PDF layout, your rule silently breaks.
- Hazel can search for text inside text-based PDFs, but it doesn't understand what a file means. It can find the string "INV-2042" if it appears in the file β but it can't tell you that the document is an invoice, who it's from, or what the amount is.
What Is FilesDesk?
FilesDesk is an AI-powered file renamer and organizer for Mac and Windows. Instead of rules, it sends each file's content to an AI model β Claude, GPT-4o, Gemini, or a local Ollama model running on your machine β and gets back a structured answer: what kind of document is this, what's the date, who's the vendor, what's the reference number?
You give FilesDesk a template. Something like {YYYY-MM-DD}_{vendor}_{type}-{number}.{ext}. It fills the slots with information it extracted from the file's actual content. The result is a clean, consistent name with no guesswork and no pattern-matching on your part.
Where FilesDesk is strong:
- Mac and Windows from a single license β the same templates and Watch Folders work on both
- Handles messy, unpredictable files: scanned receipts, screenshots, PDFs from ten different vendors with ten different layouts
- No rule writing, no maintenance when vendors change their invoices
- Optional fully offline processing via Ollama β nothing leaves your machine
- Watch Folders for continuous, set-and-forget automation
Where it's weaker:
- Not the right tool for "move file X to folder Y when condition Z is met" β that's Hazel's territory
- AI processing has a cost: cloud credits (or your own API key), or local GPU/CPU time with Ollama
- Smaller ecosystem of third-party integrations compared to Hazel's fifteen-year head start
FilesDesk vs Hazel: Head-to-Head
| Capability | Hazel | FilesDesk |
|---|---|---|
| How it decides what to do | User-written conditions + actions | AI reads file contents and structures the answer |
| Reads inside the file | Text search inside text PDFs; no semantic understanding | Yes β text, images, PDFs (including scans), screenshots, Word docs |
| Platforms | macOS only | macOS and Windows |
| Watch Folders | Yes (core feature) | Yes |
| Renaming approach | Pattern-based β needs an extractable pattern in the file | Template-based β slots filled from AI-extracted content |
| Handles unpredictable files | Rules break when layout changes | Robust β AI adapts to layout changes automatically |
| Offline / private mode | Fully local by default | Local via Ollama; cloud via Claude / GPT-4o / Gemini |
| Setup time | Hours to days for a useful rule set | Minutes β pick a template, point at a folder |
| Scripting / advanced control | Best-in-class (AppleScript, shell) | Templates + Watch Folders; no AppleScript |
| Pricing model | One-time license per machine (~$42 USD) | Free tier + lifetime license or managed credits |
| Best for | Deterministic, repeatable file flows on macOS | Messy, content-rich files across Mac and Windows |
Where the Real Differences Show Up
1. Naming files: pattern vs. content
Hazel renames using patterns. If your invoice arrives as INV-2042-AcmeCorp.pdf, Hazel can pull the invoice number and vendor with a regex and rebuild the name cleanly. That works great β until the same vendor changes their template and the file arrives as document(3).pdf with all the information buried inside the PDF.
FilesDesk skips the pattern entirely. It opens the file, reads its contents with AI, extracts the vendor, date, and reference number, and writes a name like 2026-06-20_acme_invoice-INV-2042.pdf. Same result. No regex to write, and it keeps working when Acme redesigns their invoice for the third time.
2. Sorting and moving files
Hazel wins this category and it's not close. "Move every PDF older than 30 days from Downloads to Archive" is a one-line Hazel rule. FilesDesk is built for naming and classification; raw file-management plumbing is genuinely overkill for an AI model. This is actually why the two tools work so well together β each handles the part it's built for.
3. Cross-platform support
Hazel is Mac-only and has been since day one. If you work across both operating systems β a Mac at home, a Windows machine at the office, or a team split between platforms β FilesDesk is the only tool that follows you. Same templates, same Watch Folders, both platforms.
4. Privacy
Hazel runs entirely on your Mac. Nothing goes anywhere. That matters for legal, medical, and financial files, and it's part of why the tool has such a loyal following in those industries.
FilesDesk gives you the same option via Ollama: install a local model, point FilesDesk at it, and your files never leave the machine. The cloud mode (Claude, GPT-4o, Gemini) is faster and higher quality, but the offline mode is a genuine privacy match for Hazel. See the offline AI file renamer guide for setup details.
5. Pricing
Hazel is around $42 USD as a one-time per-machine license, with a discount for additional Macs. Reasonable for what it does, and there are no ongoing fees.
FilesDesk has a free tier with trial credits so you can test it on real files before paying anything. The paid options are a lifetime license (one-time) and a managed plan that includes AI credits. If you bring your own API key or use Ollama, the ongoing AI cost goes to the provider β or is free β and you only pay the FilesDesk license once.
When to Pick Each Tool
Choose Hazel whenβ¦
- You're on Mac and plan to stay there
- Your files arrive in predictable formats with patterns you can describe as rules
- You need deterministic, scriptable, auditable automation
- You're comfortable writing and maintaining rules, and you enjoy the control that gives you
- You want a single one-time payment with no ongoing cost
Choose FilesDesk whenβ¦
- Your inbox is a mix of scans, screenshots, and invoices from different vendors in different formats
- You work on both Mac and Windows and want one consistent workflow
- You don't want to write or maintain rules β ever
- You need the AI to read what's inside a file, not just match a pattern in its name
- You handle sensitive data and want a private offline option (Ollama)
The Power-User Setup: Running Both
The cleanest insight from people who've spent time with both tools is that they're not really competing. Hazel is a routing engine. FilesDesk is a naming engine. Put them together and you have a pipeline that covers everything:
- A new file lands in your Downloads or Scanner Inbox folder.
- FilesDesk's Watch Folder catches it, reads its content with AI, and renames it to your template β
2026-06-20_acme_invoice-INV-2042.pdf. - Hazel watches the same folder, sees the now-clean and predictable name, and routes the file: invoices go to
/Finance/2026/, contracts to/Legal/Active/, receipts to/Receipts/Q2/.
FilesDesk does the reading and naming. Hazel does the routing. Each tool handles the part it's best at, and neither is doing something it wasn't built for.
The key insight: Hazel rules work beautifully on predictable file names. FilesDesk gives you predictable file names from unpredictable content. The combination means your Hazel rules can finally be simple and durable, because the input they're acting on is already clean.
Migrating from Hazel: A Practical Path
If you're moving off Hazel β usually because you've added a Windows machine or because your rules keep breaking as vendors update their templates β here's the order that causes the least disruption:
- Audit your existing rules. Group them by type: naming rules, sorting/moving rules, deletion/archiving rules.
- Replace the naming rules first. Set up a FilesDesk template that matches your convention β something like
{YYYY-MM-DD}_{vendor}_{type}-{number}. Run it on a sample folder and check the output before touching your main archive. - Leave sorting rules in Hazel for now. They're cheap to maintain, and Hazel is genuinely the best tool for that. No reason to migrate what's working.
- Point a Watch Folder at your inbox. New files will now land already named to spec. The Hazel rules that run downstream get clean, predictable input to work with.
- Optionally, bulk-rename your backlog. FilesDesk can process a folder of years-old files in one pass so your historical archive matches the same convention as everything going forward.
If you're moving to Windows entirely, FilesDesk replaces the whole stack. There's no Hazel for Windows, and File Explorer's built-in automation is nowhere near a substitute.
Stop Writing Rules. Let the AI Read the File.
FilesDesk gives you AI-powered renaming and Watch Folders on Mac and Windows. Free to try β no card required.
Download FilesDesk Free βFrequently Asked Questions
Is there a Hazel alternative for Windows?
Yes. FilesDesk runs on both Mac and Windows and is the closest equivalent to a Hazel-style Watch Folder workflow on Windows. It uses AI to read each file's content instead of rule conditions, which also removes the main reason Hazel rules break on Mac: layout changes. Hazel has no Windows version and Noodlesoft hasn't indicated one is planned.
Does FilesDesk replace Hazel completely?
Depends on what you use Hazel for. If you use Hazel mainly to rename and classify files, FilesDesk replaces that. If you use Hazel for rich conditional routing β "move this to that folder, tag it, import it to DEVONthink, then delete the original" β keep Hazel for that logic. The two are genuinely complementary.
Can Hazel read what's inside a PDF?
Hazel can search for specific text strings inside text-based PDFs. It can't interpret the document. It won't tell you "this is an invoice from Acme Corp for $1,250 dated June 3." FilesDesk can, because it sends the content to an AI model that returns structured fields β vendor, date, amount, document type β rather than a string match.
Is FilesDesk private like Hazel?
Hazel is fully local. FilesDesk can match that by using its Ollama integration β a local AI model runs on your machine, and no file data goes to any cloud service. For sensitive data, the Ollama mode is a genuine privacy equivalent to Hazel. The cloud mode (Claude, GPT-4o, Gemini) is faster and more accurate, but it's the local mode that matches Hazel's privacy profile.
How much does Hazel cost vs FilesDesk?
Hazel is around $42 USD as a one-time per-machine license. FilesDesk has a free tier with trial credits, a lifetime license option, and a managed plan with AI credits included. Bring-your-own-key users (Claude, GPT-4o, Gemini) or Ollama users pay nothing to FilesDesk for AI processing beyond the base license.
Can I use Hazel and FilesDesk together?
Yes β and this is actually the setup most power users land on. FilesDesk watches the inbox and names incoming files from their content. Hazel watches the same folder and routes the now-predictable names to the right archive location. FilesDesk names, Hazel routes.
What is the best file automation tool for Mac in 2026?
For deterministic, rule-based automation on predictable files, Hazel is still unmatched. For content-based renaming of messy or varied files, FilesDesk is the better choice. If you work across Mac and Windows, FilesDesk is the only option that follows you. The honest answer for most people with a mixed inbox is to use both.
The Bottom Line
Hazel is still the best rules engine for macOS, full stop. Fifteen years of refinement, a loyal community, and deep AppleScript integration make it the right tool when your files behave and your platform is Mac.
FilesDesk is the right tool when your files don't behave β or when your platform is Windows, or both. It doesn't try to replace Hazel's routing logic. It reads what Hazel can't: the content inside a file, not just the pattern on its surface.
If your files are predictable and your Mac setup is locked in, Hazel will serve you well for another decade. If your inbox is a pile of scanned PDFs, mixed-vendor invoices, and screenshots with names like document(3).pdf β or if you've got a Windows machine in the mix β FilesDesk is the upgrade. And if you run both? That's probably the sharpest setup of the two.
Further reading: file naming conventions for Mac & Windows, setting up Watch Folders for automatic renaming, AI file organizer for Mac, and offline AI renaming with Ollama.