Rectangle alternatives: decide what you can drop

Looking for a replacement window manager usually starts in the wrong place. The instinct is to open a list of alternatives and compare them, but window managers on macOS overlap heavily, and comparison lists are ordered by how many features each one has rather than by which features anyone actually uses. The result is a long read that ends with the same question it started with.

A faster route is to decide what can be dropped first. Almost everyone in this category uses two or three behaviours daily and has the rest switched on out of habit. Once the short list of daily behaviours is written down, most candidates disqualify themselves in a minute, and price stops being the deciding factor.

Split the feature set into eight axes

Every tool in this category is some subset of the following. Count how many of these were used in the last week, honestly.

• Keyboard shortcuts for halves, quarters, and maximise
• Drag-to-edge and drag-to-corner snapping
• Fine grids: thirds, sixths, ninths
• Moving windows between multiple displays
• Configurable gaps between windows or at screen edges
• Pinning one app to a side of the screen for a working layout
• Choosing actions from a menu bar menu
• Exporting and importing the configuration

Most people land on the first two plus display movement. That answer opens the field wide, including options that cost nothing. A smaller group uses gaps or the pinned sidebar every day, and that narrows the field to two or three products immediately. Either way the count takes two minutes and removes most of the reading.

Check what built-in macOS tiling already covers

Current macOS versions tile windows without any third-party help. The settings live in System Settings under Desktop & Dock, and they are stored as discrete preferences in the window manager's own domain: whether dragging to a screen edge tiles a window, whether dragging to the top edge behaves differently, whether tiled windows get margins between them, and whether a modifier key is required. On a machine running macOS 26.6.2 those keys are present and individually switchable.

The dividing line is clear enough to put in a table.

What you want	Built into macOS	Third-party tool
Left and right halves	Yes	Yes
Quarters in the corners	Yes	Yes
Thirds and sixths	No	Product dependent
Arbitrary keyboard bindings	Limited	Yes
Numeric gaps at screen edges	Margins on or off only	Available in some
Pin one app to a side	No	Product dependent

If the answer to the counting exercise was halves and quarters, the honest conclusion is that no replacement is needed. That outcome has a second benefit: one fewer resident app, and one more free slot in the menu bar. Plenty of searches for an alternative end here once the built-in behaviour is actually tried for a week.

What the candidates cost, as of 10 September 2026

Prices below were read from each vendor on 10 September 2026, in United States dollars. Regional pricing and tax treatment vary, so check the current figure before buying.

Candidate	Where	Price	Requires	Last updated
Rectangle	rectangleapp.com, Homebrew	Free, MIT licensed	macOS 10.15+	v1.100, 26 Aug 2026
Rectangle Pro	rectangleapp.com via Paddle	$9.99 one time	Stated on site	Actively developed
Magnet	Mac App Store	$4.99	macOS 13.0+	12 Apr 2025
BetterSnapTool	Mac App Store	$1.99	macOS 10.12+	5 Sep 2024
Moom Classic	Mac App Store	$6.99	macOS 10.13+	6 Jul 2026
Moom (direct)	manytricks.com	$15, upgrade $8	macOS 10.13+	Actively developed
Loop	GitHub	Free, open source	macOS 13+	Actively developed

A few notes that the table cannot hold.

Rectangle Pro is a one-time purchase with a ten-day trial, and the vendor states that one licence can be active on three devices at a time. It can load shortcuts from the free Rectangle, which removes most of the setup work for anyone already using it. It is built by the same developer, on top of the same codebase.

Moom is the case most likely to cause confusion, because two products share the name. The Mac App Store listing is called Moom Classic at $6.99. The direct version from Many Tricks is $15, with an $8 upgrade path for previous buyers, and the vendor states it includes at least one year of updates while the licence itself remains valid forever. Comparing on price alone will land you on a different product than expected.

Rectangle itself is MIT licensed, and its repository shows the code descends from Spectacle. Anyone who wants to audit what a window manager does with accessibility permission can read it, which is not true of most of this list.

Price is not the thing that goes stale

Comparison articles rot because the assumptions change, not because the numbers do. Once a shortlist exists, four things are worth verifying directly.

Is it still maintained. This category has well-known apps that are still downloadable while development has stopped, sometimes with the author saying so plainly.

This project is not being actively maintained. Spectacle will remain available for download but please use at your own risk. Source: github.com (checked 10 September 2026)

That note belongs to Spectacle, the project Rectangle was built from. The same page points readers to Rectangle as the open source successor. Spectacle can still be downloaded today, which is exactly why it keeps showing up in search results. Being available and being a sensible thing to start using in 2026 are different questions.

Is it still being sold to new customers. Some apps continue to work for existing licence holders while new sign-ups are closed.

Has the owner changed. This has happened nearby: the best-known menu bar organizer changed hands in 2024, which became a standard thing for buyers to check. Ownership changes are neither good nor bad on their own, but they change who is answering support mail.

Has the pricing model changed. Several Mac utilities have moved from a plain one-time purchase to a purchase that includes a fixed window of updates. Moom's direct edition is written that way today.

Last-updated dates help with all four. Two apps in the table above are separated by nearly two years on that column. A gap is not a defect, but it is the only public signal for how a product will respond when the next macOS release changes something.

Free and open source is not one category

Three entries in the table cost nothing, and they are not interchangeable. The distinction that matters when picking one is who is on the hook when it breaks.

Rectangle is MIT licensed, has been developed since 2019, ships signed builds from a website, and updates itself through Sparkle rather than requiring a manual download each time. It also has a commercial sibling, which is a reasonable proxy for the project continuing to be worth someone's time.

Loop takes a different shape. It is open source, requires macOS 13 or later, and its interaction model is a radial menu triggered by holding a key rather than a set of individual shortcuts. Anyone whose hands already know a shortcut scheme will find that a bigger change than the price suggests, and anyone who finds shortcut schemes hard to memorise may find it a smaller one.

Then there are the tiling window managers driven by configuration files rather than settings windows. They are free, powerful, and a different category of commitment. The relevant question is not whether they are capable, it is whether editing a config file to change a layout sounds like a Saturday well spent.

The general point holds beyond this list. Free means no invoice, not no cost. The cost shows up as setup time, as the risk of the project going quiet, and as the effort of learning an interaction model that differs from the one currently in muscle memory. Weigh that against a one-time purchase of a few dollars before assuming the free option is automatically the cheaper one.

Moving your configuration is the real switching cost

The overlooked expense in switching is rebuilding shortcuts. Rectangle exports and imports its configuration as JSON from the Settings window, stores preferences at ~/Library/Preferences/com.knollsoft.Rectangle.plist, and will read a config file dropped at a known path on launch, renaming it afterwards so it is not read twice.

None of that transfers to another vendor. Formats are product specific, so switching means entering bindings by hand. Count the bindings currently in use before starting: someone with four will finish in five minutes, someone with twenty five is looking at a real afternoon. That number, not the purchase price, is usually the larger cost.

There is also a version caveat worth knowing if you are rolling back rather than moving forward. Rectangle changed how shortcuts are stored at v0.41, and configurations written since then will not load into builds older than that.

Trying candidates without making a mess

Every app in this category needs accessibility permission, because moving another app's windows is impossible without it. Install three candidates and the permission list holds three entries, and it stops being obvious which one is doing what.

Two rules keep the trial clean. Run at most two at once, because three window managers competing for the same key combinations makes it impossible to tell which product is producing which behaviour. And when removing one, take it out of the accessibility list as well as the Applications folder, since a stale entry makes the next diagnosis harder.

Installing through Homebrew helps here too: uninstalling with the zap flag clears the stored preferences, whereas a manual drag to the trash leaves them behind, so reinstalling later resurrects the old state.

The menu bar cost of a trial

There is one more expense that never appears in comparison tables. These are all resident apps, and each one places an icon. During a trial, the old app and the new one are both installed, so two slots are occupied at once on a bar that was already full.

That matters for the quality of the comparison, not just for tidiness. With both icons visible it is easy to click the wrong one and misattribute what happened. Collapsing one of them keeps it running and reachable while leaving a single icon on display, which makes the two weeks of testing actually comparable. What a menu bar organizer can reach while collapsed varies by product, and the range is described on features, with the differences in approach on how it compares.

What to change first

Spend one week on built-in macOS tiling alone and write down every action that was missing. That list, not a comparison table, is the shortlist. Then trial at most two candidates against it, and if the bar is already crowded, look at Koffret before adding a second icon to it.