The Arc browser, even when it was in beta back in 2023, struck a chord with me and many of its early users. Certainly, it was designed for power users: robust tab management (Split views + folders), workflow- and cookie-segmented sidebars (Spaces), custom site-specific CSS and JavaScript tuning (Boosts). These features were available out of the box, without having to juggle browser extensions and add-ons. But despite the strong feature set, Arc also felt well-designed[1]; its developers were mindful of its user interactions, and compared to its competitors like Vivaldi, did not feel as overwhelming[2].
The company behind the browser is the aptly named The Browser Company (TBC). They were venture-backed startup, with a sizable team of ~80 employees when their flagship product wasn't even formally launched; all that delicate design and engineering took a lot of development effort. At the time, I was already concerned about the sustainability of their business, as companies prove time and again that consumers are a finicky bunch.
TBC likely saw the same limitations, and in 2025 shifted their attention entirely to a new platform, the AI-focused and -integrated Dia browser. Arc was demoted to regular Chromium updates only, which sustained a shrinking group of passionate enthusiasts in the short term while functionality continued to degrade. The Windows port developed major memory issues, while the macOS version stopped playing videos in full screen after a routine update. It speaks to just how good the product is that we Arc diehards stuck with it, even with the explicit deprecation and the resulting thousand papercuts.
Alas, the end is here. TBC, now a part of Atlassian, announced that they will be sunsetting Arc accounts by late December, with account data deleted after 30 days. Presumably, the browser will then finally shut down[3].
I am completely invested in this saga, as the community is now forced to figure out exit strategies. Sure, there were the initial phases of anger, denial and pleading, but very quickly Arc users accepted that the 1½ years of slow degradation had reached its natural conclusion, and started looking for alternatives. The announcement has produced, and resurfaced, a dozen Arc lookalikes and clones with a mix of features and development philosophies.
Currently, the most mature successor is the Zen browser. It was launched in 2024, with a similar design language to Arc but with Gecko—the web engine developed by Mozilla—instead of Chromium as its renderer. Beyond the look-and-feel, however, Zen's management directly contrasts with TBC's: Zen is open source, run by its small community of contributors. While it is not yet at feature parity to Arc, its developers have welcomed Arc refugees with open arms[4].
Of the rest of the Arc-likes, most look to be smaller efforts, run by single developers with telltale signs of AI-assisted development. I've checked out a couple of them, and apart from capitalizing on the shutdown announcement, they show little product differentiation. The challenge, and biggest shortcoming, in this new set of experimental browsers is the risk of not following through—developers get bored, or they don't see traction, or they run into the same monetization roadblock that doomed Arc, and their project ends up in the same predicament.
And small development teams cut both ways. They avoid the high overhead of TBC's full team, with less pressure to grow user counts or produce revenues or advance a business model[5]. Cloning Arc's features is a straightforward task for AI coding agents—that's how so many alternatives have already popped up—which enable solo efforts to go much farther. On the flip side, AI doesn't solve the problem of garnering user trust, particularly in the context of web browsers designed to run webapps with user data.
Right now, of all the available options, I'm trying out Crest Browser. It has implemented most of Arc's features and fixed some annoying bugs that have cropped up during its dormancy, with the twist that the developer does not want to copy Arc outright. Indeed, its namesake refers to a sophisticated medieval crest system for its iconography. The configurator comes complete with emblems and bands and geometric shapes, and whose aesthetic immediately turns off some users.
So you know the project is handled with conviction and attitude. In the otherwise staid world of web browsers, that's a standout proposition.
It also has one of the most over-the-top theme customizers I've seen. ↩︎
Sadly, one of the reasons for Arc's demise was that it was still confusing to regular, non-power users, who didn't use most of its best features. ↩︎
I have seen one enterprising engineer spoof the Arc account server to try to keep it alive, but at this point nobody knows whether that'd even work. ↩︎
I tried Zen for a few weeks. There's a lot to like about the browser, but unfortunately Gecko is a distant third web rendering engine, which means it's not a priority for web developers to test and optimize for. The Google Workspace suite, Docs and Sheets in particular, runs poorly in Zen, with annoying rendering bugs. ↩︎
Worryingly, my chosen MacOS Spotlight replacement, Raycast, is falling into the same pattern: great free features at launch, but is increasingly pushing monetization via AI-powered features. ↩︎
