Arcの代わり: what it does and where it breaks down

Searching for Arcの代わり turns up a long list of candidates, and the list is not the problem. The problem is that the phrase points at three different kinds of software at once. One kind restores the look and the keyboard speed of a sidebar full of named tab groups. Another moves the daily web apps out of tabs entirely and gives each one a persistent pane with its own session. A third does nothing new at all and just uses the profile feature already sitting in a mainstream browser. These three shapes solve different problems, and picking the wrong shape is the reason a replacement can feel fine in a demo and wrong after a week.

What follows separates the shapes, states what each one does and does not do, and names the point where each tends to break down.

Three products answer to the same search

Shape A is a replacement for the interface. Vertical tabs in a sidebar, named and colored sections, fast keyboard movement between them, a calm window. Most Chromium based browsers competing in this space sit here. The thing being organized is what the eye sees.

Shape B is a replacement for where the apps live. Gmail, Slack, a project tool, a calendar, a support inbox: instead of being tabs that get lost, each one becomes a persistent pane, usually with its own cookie jar. The category name for this is an app aggregation browser. The thing being organized is the session and the notification, not the tab strip.

Shape C is not a new product at all. It is the profile feature in a mainstream browser, used deliberately. Two profiles, two sets of cookies, two dock icons. Cost is zero, and separation is complete.

These shapes are not exclusive. Plenty of setups run Shape C for identity and Shape A for daily browsing, or keep a normal browser for research and put only the permanent work apps into Shape B. Trouble starts when someone skips the question of which shape applies and starts installing whatever ranks highest in a list of alternatives. Swapping one Shape A product for another Shape A product will never fix a complaint about accounts bleeding into each other, because no amount of tab organization touches a cookie jar.

What Arc is today, stated plainly

The reason people start looking is usually something they read rather than something that broke. The facts are on the record, so here they are without interpretation.

Arc's own download page carries a notice about the state of the build:

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

In a letter published on May 26, 2025, the maker stated that active development of new features had stopped, while Chromium upgrades and security fixes continued. The same letter explained that open sourcing the browser is complicated by an internal framework shared with the newer product, and that shutting Arc down was not the intention.

On September 4, 2025, Atlassian announced an agreement to acquire the company behind both browsers. That announcement is about the newer AI focused browser, and it makes no new commitments about Arc.

Two conclusions follow, and only two. Arc still runs. Arc is not going to grow. Anyone evaluating a replacement in 2026 is doing so because the roadmap ended, not because the software failed.

What each shape actually separates

Shape	Boundary it draws	Strong at	Does not do
A, vertical tab browser	Visual grouping	Tab order, keyboard navigation, a quieter window	Keep two accounts of one service signed in at once
B, app aggregation browser	App placement and session	Returning to permanent apps, notifications, parallel accounts	Handle piles of throwaway research tabs
C, browser profiles	Cookie storage	Complete account isolation at no cost	Fast switching, one glance at everything

The right hand column decides most cases. Shape A makes the screen calm within an hour of setup, which is why it demos so well. It does not give a second Gmail account anywhere to live. Two named sections inside one profile share one login, and signing out of one signs out of the other.

Shape B usually gives each pane its own storage, so three accounts of the same service can sit side by side and stay signed in. How far that isolation goes is an implementation detail that varies by product, so the fastest check is to read how a candidate describes its Workspaces and see whether the description talks about sessions or only about layout. A product whose section feature is described purely in visual terms belongs in Shape A, whatever the marketing calls it.

The login test that removes half the candidates

There is one test, it takes two minutes, and it settles the shape question.

Open the same service twice with two different accounts in the environment being evaluated. If the second sign in knocks out the first, the boundary is cosmetic. If both stay signed in and both keep receiving notifications, the boundary is real.

For anyone running more than one identity, the ability to hold a separate workspace for each app is not a nice extra. It is the entire requirement, and every product that fails the test can be dropped without further reading. For anyone with exactly one account per service, the test is irrelevant and the decision moves to speed, memory use, and how the window feels at the end of a long day.

Shape C passes the test automatically, because profiles are cookie jars by definition, and it costs nothing. The reason people still go looking is friction. Switching profiles means switching windows, and every switch costs a second of attention plus the moment spent finding the right window. At fifty switches a day that friction is the whole problem, not a detail.

Where each shape breaks down

Shape A breaks down after the move, not during it. Extensions that were load bearing may behave differently. Some streaming services are picky about the build. The default browser setting on macOS needs attention. The keyboard shortcuts built up over several years have to be relearned, and for the first few days the new setup is measurably slower than the old one. None of this is fatal, but it all lands in the same week.

Shape B breaks down when the working style does not match the design. A tool built around permanent panes is not built for opening forty search results, skimming them, and closing them. It is also only as good as its coverage: a service that is not supported gets added as a generic page, which often means weaker notification handling. Checking a candidate's list of supported apps against the services opened every single day takes five minutes and prevents most of the disappointment.

Shape C breaks down at scale. Two profiles is comfortable. Past five, nobody can remember which window holds which identity, and the failure mode is sending something from the wrong account. Whether the windows can be visually distinguished by color or name is what sets the practical ceiling.

What the usage numbers suggest

The letter from Arc's maker included figures about how the browser was actually used, and those figures are more useful for choosing a replacement than any feature list.

Only 5.52% of DAUs use more than one Space regularly. Only 4.17% use Live Folders (including GitHub Live Folders). Source: browsercompany.substack.com

Just 5.52% of daily users regularly kept more than one named section. For everyone else, the value was the calm window and the speed, not the sectioning feature at all.

That reframes the decision. Someone who lived in a single section and is now shopping for a product that advertises unlimited sections is buying a solution to a problem they never had. What they want back is the interface, which puts them in Shape A, and the cheapest good answer there is often a mainstream browser configured carefully.

Someone in the remaining few percent is in a different position. Real separation, persistent sign ins, notifications arriving without the window being focused: that is Shape B, and it usually costs money. Reading how a candidate structures its pricing, and specifically whether the free tier is capped by app count or by feature, saves more time than reading reviews. For narrowing between specific products, a one to one comparison such as Compared with Wavebox gives sharper answers than a ranked list of ten.

Research and residency do not belong in the same tool

Part of what makes the decision hard is an assumption nobody states out loud: that one browser has to hold everything. Two very different activities are being asked to share a window.

The first is residency. Mail, chat, calendar, task tracker, a support queue. These stay open, get returned to dozens of times a day, and are almost never closed. There are maybe eight of them, and they are the same eight every week.

The second is research. Search, read, maybe save, close. This produces far more tabs than residency ever will, and almost all of them are disposable within an hour.

Forcing both into one tool makes one of them worse. Pour research into something designed around permanent panes and the careful layout fills with debris. Put residency into an ordinary browser and the permanent apps sink into the same tab strip as everything else, which is the original complaint. Arc was liked partly because pinned tabs and normal tabs gave these two activities different treatment inside one window.

A replacement does not have to preserve that arrangement in a single application. Splitting by role is legitimate and often easier: keep the familiar browser for research, and move only the permanent apps somewhere built for them. Checking what a candidate says it can host, on its own list of Features, is enough to decide whether the residency half fits. This split also makes migration reversible, since the research half never moves and therefore never has to be moved back.

Count four things before installing anything

Before downloading a single candidate, write down four numbers. The list of candidates shortens itself.

• The number of web apps opened every day without exception. Above ten, free tier caps become the deciding factor.
• The number of accounts held on the same service. Two or more makes session isolation mandatory.
• The names of the extensions that cannot be given up. A long list restricts the field to Chromium based products.
• The number of times per day attention stops while hunting for a window. A rough estimate is fine.

One more thing is worth writing down alongside them: the keyboard shortcuts currently in use for switching tabs, moving between sections, toggling the sidebar, and jumping to the address field. When a new tool assigns different keys to those four actions, the first days produce constant misfires, because the hands move before the thought does. Most products allow the bindings to be changed, so remapping them to the old layout on day one roughly halves the adjustment period. Skipping that step is a common reason a perfectly reasonable tool gets judged unusable in week one.

These four numbers answer the shape question faster than a week of trials. Without them, every candidate produces the same verdict, which is that it seems fine but nothing about it is compelling, and the trial ends back where it started.

What to change first

Pick the shape before picking the product, using the login test and the four numbers above. Then move exactly three daily apps into the new setup and leave everything else alone for a week, because a partial move is reversible and a full move is not. If session isolation turned out to be the requirement, start with a tool built around it, such as SpaceDeck, and add the rest only after the first three have survived a normal week.