arcvsdia

what each one actually is

Arc is the browser the browser company built between 2022 and 2024, now scheduled for shutdown on 22 december 2026. it runs on chromium and introduced a fundamentally different spatial model for browsing: sidebar tabs organized into spaces, with a command bar (⌘T), little arc for quickly opening and closing URLs, and a design language that felt more like a Mac app than a browser. it earned an unusually devoted user base and changed how a lot of people thought about browser UX. development stopped in may 2025, and the browser company itself was acquired by atlassian for $610 million later that year — arc sat in maintenance mode from then until october 2026, when it was given a shutdown date.

Dia is the browser company's flagship product, and now atlassian's bet on an ai-first browser following the acquisition. it's also chromium-based, but without arc's tab/space model. dia's core proposition is an AI assistant that lives inside the browser: summarize the page you're on, ask questions about a document, manage tabs intelligently, extract information from multiple sites. the UI is simpler and more conventional than arc's, which is either a relief or a regression depending on what you loved about arc.

pricing, honestly

arc has always been free, and the price has stopped being the relevant fact. it is being switched off on 22 december 2026. between now and then it still runs, but new accounts can no longer be created and the mobile apps lose ai features and sync on 23 november.

dia's free tier gives you full browsing with a capped number of AI interactions. dia pro is $20/month and lifts the AI limits entirely; in august 2026 a $100/month tier arrived above it for heavy use. the AI features — tab summarization, page Q&A, skills, intelligent session management — are what you're paying for; core browsing stays free. whether that justifies $20 depends entirely on whether you use AI browsing assistance daily or find it a novelty.

one hard constraint that now really matters: dia runs only on macos 14+ on apple silicon. if you are on windows or an intel mac, dia is not available to you and arc stops working on 22 december regardless — so your migration is to chrome, firefox, brave or safari, not to dia. the browser company has no answer for you here.

what it's actually like to use them

arc's UX was distinctive and required an onboarding mindset shift. once you internalized spaces, the sidebar, and the distinction between pinned and unpinned tabs, it clicked. the command bar made navigation fast. the design was opinionated in ways that paid off after a week. the main friction: it's heavier on RAM than chrome and occasionally crashed in ways chromium browsers shouldn't.

dia is easier to pick up. if you've used a conventional browser, the basics work as expected. the AI features are accessible — a small button opens the AI panel, you can highlight text and ask about it, tabs are shown simply. what dia hasn't yet cracked is the power-user density arc had. arc's spaces gave you cognitive control over complex projects. dia doesn't have an equivalent yet. if you managed five active projects in arc through spaces, dia will feel like a step backward in organization until (or unless) the AI catches up to doing that automatically.

who dia is for

• anyone starting with the browser company's products in 2026
• users who found arc overwhelming or whose team doesn't want to onboard to arc's model
• people who actively use AI for research, note-taking, and reading — dia's AI is genuinely useful
• anyone who wants to be on the product that's actively receiving new features and fixes

who arc is still fine for

• arc loyalists who've built elaborate spaces and workflows they're not ready to abandon
• users where arc's tab model is exactly what they need and the AI features aren't a priority
• anyone who wants to evaluate arc before dia's AI matures further

when to avoid each

you cannot start on arc. new account creation stopped on 22 october 2026 and the browser stops working on 22 december. this is not a judgement call any more.

don't expect dia to replicate the organizational power of arc's spaces out of the box. if your workflow depended on the spatial tab model, dia will feel sparse until you find new patterns (or the AI gets good enough to replace the manual organization).

stuff their landing pages won't tell you

• arc syncs tabs across devices via iCloud; dia uses its own sync — the experience is different
• arc is mac and iOS only; dia is expected to expand to more platforms than arc
• the browser company's public announcement about arc's future was unusually direct for a startup — take it at face value, and the subsequent atlassian acquisition confirmed it wasn't a bluff
• dia is macos-only and needs apple silicon; windows and intel mac users have no migration path from arc to dia at all
• dia's pricing is still moving — the company has said it wants a wider range of tiers eventually, from around $5/month up to enterprise
• arc had a windows version in limited beta; dia's windows availability has not been confirmed at the same pace
• both browsers are chromium — your extensions work in both without modification

the call

dia. arc is finished in the sense that matters most: it's not where the company is shipping, and the company itself now belongs to atlassian. the browser company's talent, product roadmap, and backing are behind dia now. picking arc in 2026 is picking a browser that will slowly fall further behind rather than further ahead.

if you loved arc enough that you want to wait and see whether dia gets good enough to replace it, that's a reasonable position. use arc while it works, evaluate dia periodically. but if you're starting fresh: dia is where you should land.