Arcの代わり alternatives: what you can drop

The replacement search usually stalls in the same place. A list of alternatives gets opened, three of them get installed over a weekend, and by Monday the old browser is back in the Dock. Nothing on the list was bad. The problem is the comparison itself. Arc is not one feature, so nothing else is going to be one-to-one with it. It is roughly a dozen separate behaviors that happen to ship in the same application, and for any given person only three or four of them carry real weight. The rest can go without being noticed. The useful question is not which product is closest to Arc. It is which behaviors would actually be missed on day two, and which ones are already dead weight.

Arc is a bundle, and that is why nothing matches it

Write out what is actually in the app and the size of the problem changes. Spaces. Profiles. The sidebar with vertical tabs. Pinned and favorited tabs. Automatic archiving of tabs that have gone stale. Split View. The command bar. Little Arc windows for one-off links. Boosts for changing how a site looks. Easels. Themes. Arc Max for the AI features. Arc Search on iPhone. Sync across machines.

That is fourteen things. A candidate that covers ten of them reads as a failure, because the four it misses are the four that get tested first. Meanwhile a candidate that covers four could be the correct answer, if they are the right four.

This is also why recommendation threads disagree so violently. Two people who both say they miss Arc are often missing completely different parts of it. One misses keeping a work Google account and a personal one signed in side by side. Another misses vertical tabs. Another misses the command bar and would be satisfied by a launcher that is not a browser at all. None of them are wrong, and none of their answers transfer.

The practical move is to stop evaluating whole products for a while. Break the bundle into parts, find out which parts get used, and shop only for those. A shorter list of requirements makes the choice much easier, and it makes the result stick, because it is no longer a compromise against a feature list nobody was using.

What the official pages say right now

Two facts set the shape of the decision, and both are stated publicly rather than inferred.

The first is on the front page of arc.net itself, presented as a notice next to the download buttons.

FYI: Arc receives Chromium updates only. For active security patches and enterprise-grade protection, download Dia instead. Source: arc.net

The second is ownership. The Browser Company of New York, which builds Arc and Dia, was acquired by Atlassian in a cash deal of 610 million dollars, announced on 4 September 2025 and completed on 21 October 2025. Development attention sits with Dia, which now runs a free tier plus paid plans at 20 dollars and 100 dollars per month, with a 14 day trial, as listed on the Dia plans page.

Neither fact forces anyone to move today. Arc still downloads, still runs, and still gets the Chromium security work described in that notice. What both facts do is remove one specific hope: that the missing pieces will be added later. Whatever is in the app today is what there is. That makes an inventory worth doing now rather than after something breaks.

Take the inventory before shopping

Five working days of notes are enough, and the notes do not need a tool. A plain text file with three columns does the job: what got switched to, why, and whether a window for it was already open somewhere.

Four numbers come out of that file, and each one points at a different requirement.

How many times a link opened under the wrong account. This is the one that quietly costs the most. If it happens more than twice a day, isolation of logins is the load bearing part, and tab organization is not. Someone with this number high will be unhappy in any tool that keeps one cookie jar, no matter how good its tab management is.

How many times a Space got switched. Under about ten a day, Spaces are a filing habit rather than a workflow. Over about thirty, the switch itself is the thing to optimise, and how fast it happens matters more than how it looks.

How many times a tab got hunted for. High numbers here point at search and at pinning, not at grouping.

How many features got used at all. Most people find that Boosts, Easels and themes appear zero times across the whole week. That is not a criticism of those features. It is a measurement, and it is the measurement that makes the next two sections possible.

The rule for reading the file is blunt. A behavior used dozens of times a day is expensive to lose, because the cost is paid in interruptions all day long. A behavior used once a week is cheap to lose, because a workaround only has to be tolerable once a week.

The parts that are cheap to drop

Four categories tend to fall out of the list without pain.

Visual customisation goes first. Themes and per-site restyling get set up once, admired, and then never touched. Nothing downstream depends on them.

The canvas and notes features go next. Easels are pleasant, but almost everyone who uses them has a notes app open anyway, and the work usually ends up there in the end.

One-off link windows come third. A quick throwaway window for a link from Slack is genuinely useful, but a second browser on the machine covers the same need. Most Macs already have Safari installed and signed out, which makes it a serviceable scratch window.

The AI layer is fourth, and it is the one people hesitate over. The hesitation is worth examining. AI that lives inside the browser chrome is convenient, but the same models are reachable from a separate window or an extension, and the pricing moves independently of the browser. Treating it as a separate purchase decision rather than a browser feature keeps both decisions cleaner.

The test to apply is the same every time. Open the feature, try to remember the last time it was used deliberately, and if the answer is longer ago than a week, cross it off. Dropping it costs nothing, and every item crossed off widens the pool of acceptable replacements.

The parts that are expensive to drop

The remaining list is usually short, and it is usually some subset of five things.

Separate sessions for the same service. Two Google accounts, two Slack workspaces, two Notion logins, all signed in at once with no picker dialog. This is the one that no amount of tab tidiness substitutes for.

Switching without a modal. Whatever the unit is, getting to it should be one keystroke or one click, not a menu.

Two things visible at once. A document on the left and a call or a spec on the right. Losing this pushes the work onto window management, which is slower.

Tabs that stay put. A pinned set that survives a restart, in the same order, per context.

Keyboard-first navigation. Reaching any page or app without the trackpad.

Those five are used constantly, which is exactly why they are worth paying for in whatever form. Everything else on the original fourteen is negotiable.

Some features were covering for a habit

A few items on the list are harder to classify, because they were never doing work on their own. They were compensating for something else.

Automatic archiving is the clearest case. It exists because tabs get opened and never closed. The feature removes the evidence rather than the cause, and it does so quietly enough that the habit never has to change. Dropping it feels alarming for about three days, after which most people simply close things. Anyone who moves to a tool without it and then complains about clutter is usually looking at a habit, not a missing feature.

Pinned tab sprawl works the same way. A pinned strip that has grown to twenty items is not a workspace, it is a bookmark bar with worse ergonomics. Rebuilding it from scratch in a new tool, and refusing to pin anything until it has been opened twice in a week, usually lands somewhere between five and eight items. That smaller strip is faster to scan and survives a move to almost anything.

The sidebar itself can fall into this category. For some people it is the map of the day. For others it is a place where reading material accumulates so it does not have to be dealt with. The two look identical in a screenshot and behave completely differently in practice.

This matters for the replacement decision because habit-covering features distort the shortlist. They make a candidate look inadequate when the real gap is on the user's side. Separating the two before shopping means fewer requirements, a shorter list, and no money spent on a feature whose job was to hide a pile of unread pages.

Where each surviving part can come from

Once the list is short, the sourcing is straightforward, and it does not have to come from a single application.

Behavior kept	Where it can come from	What it costs
Separate logins, same service	Browser profiles in Chrome or Edge, Safari profiles, or a browser that gives each app its own window	Free in the built-in options, paid in most dedicated tools
Grouping by project or client	Tab groups, or per-app workspaces	Free to try in every candidate
Fast switching	Keyboard shortcuts, or a dock of apps down one side	Depends on the tool, worth testing on day one
Split view	The macOS window tiling built into recent releases, or split panes inside the browser	Free at the OS level
Vertical tabs	Edge, Vivaldi, Zen, Orion, and most app aggregation browsers	Free
Keyboard-first launching	Spotlight, or a third party launcher, independent of the browser	Free or one purchase

Two notes on reading that table. First, the built-in options are free and already installed, which makes them the correct baseline to measure paid tools against. Apple documents profiles and window tiling in its own macOS support material, and Google documents Chrome profiles in the Chrome help pages, so the baseline can be set up in an afternoon without spending anything.

Second, the paid tools in this category are not selling any single row of that table. They are selling all of them arriving pre-assembled, plus the per-app window model that profiles do not give. Whether that is worth a subscription depends entirely on the four numbers from the inventory. A reader whose wrong-account count was zero should probably not be paying for this category at all. A reader whose count was fifteen will get the money back in a week.

For the per-app model specifically, the shape of it is easier to judge by looking at how Workspaces are structured and at which services are already covered on the Supported apps list, before installing anything. Pricing across this category varies widely, and comparing it against the Pricing of a like-for-like tool is a fair way to see what the going rate actually is.

What to change first

Run the five day inventory before installing anything, then cross off every behavior that scored zero. If the wrong-account count was the number that stood out, fix isolation first and leave tab organization alone entirely, since one change to how logins are separated usually makes the tab problem smaller on its own. A per-app window model is one way to get that, and SpaceDeck is built around it, though browser profiles cover the same ground for free and are the right thing to try first.