Taste and craft. Early adopters loved the Nintendo/Disney-like product feel Miller described (Josh Miller / The Browser Company). Strong D1 for a niche. Miller writes day-one retention was strong and stayers were "fanatics," with metrics more like a specialized pro tool than mass consumer software (Josh Miller / The Browser Company).

Arc: beloved browser, novelty tax

Q: What can indie builders learn from Arc?

Measure revealed preference. Miller cites that only 5.52% of DAUs use more than one Space regularly, 4.17% use Live Folders, and 0.4% use Calendar Preview on Hover — vs Dia features used by ~40% and ~37% of DAUs (Josh Miller / The Browser Company). Switching costs for browsers are high; novelty without immediate value is a tax, not a moat (Josh Miller / The Browser Company). If the mission is mass-market infrastructure, a beloved niche product can still be the wrong vehicle — Miller says he would have stopped Arc a year earlier (Josh Miller / The Browser Company).

Q: What didn't work for Arc, and what are the risks?

Novelty tax and complexity. "For most people, Arc was simply too different, with too many new things to learn, for too little reward" (Josh Miller / The Browser Company). Not open-sourcing. Fans asked; Miller says Arc depends on proprietary ADK, also core to Dia, so open-sourcing Arc would give away Dia's IP (Josh Miller / The Browser Company, TechCrunch). User trust hit. Sunset after a pivot and acquisition left many users disappointed on social media, per TechCrunch (TechCrunch). Engagement percentages are company-reported in Miller's letter.

Key takeaways

Arc grew a passionate user base, but Miller later said most people faced a 'novelty tax': too different for too little reward.
May 2025 letter: the company stopped building new Arc features and bet on Dia; Miller writes he should have stopped Arc a year earlier.
October 2026: Arc scheduled to shut down on Mac/Windows on 22 Dec 2026 after Atlassian acquired The Browser Company for $610M in 2025.

What it is

Arc was a Chromium-based desktop browser from The Browser Company of New York, known for Spaces, vertical tabs and opinionated UI. CEO Josh Miller framed the original mission as building an "Internet Computer" — a browser that felt like home (Josh Miller / The Browser Company).

Timeline

2019 idea / early 2020s build: company starts from the belief that the browser is the real desktop OS and is under-loved (Josh Miller / The Browser Company).
May 2025: Miller's "Letter to Arc members" explains the pivot to Dia, an AI-first browser; Arc goes to maintenance (Chromium upgrades and security) without new feature development (Josh Miller / The Browser Company, TechCrunch).
2025: Atlassian acquires The Browser Company for $610M, per TechCrunch (TechCrunch).
22 Oct 2026: company announces Arc will stop working on Mac and Windows on 22 Dec 2026; new accounts already closed; Arc Search AI features wind down earlier (TechCrunch).

Why it worked (at first)

Taste and craft. Early adopters loved the Nintendo/Disney-like product feel Miller described (Josh Miller / The Browser Company).
Strong D1 for a niche. Miller writes day-one retention was strong and stayers were "fanatics," with metrics more like a specialized pro tool than mass consumer software (Josh Miller / The Browser Company).

What to learn

Measure revealed preference. Miller cites that only 5.52% of DAUs use more than one Space regularly, 4.17% use Live Folders, and 0.4% use Calendar Preview on Hover — vs Dia features used by ~40% and ~37% of DAUs (Josh Miller / The Browser Company).
Switching costs for browsers are high; novelty without immediate value is a tax, not a moat (Josh Miller / The Browser Company).
If the mission is mass-market infrastructure, a beloved niche product can still be the wrong vehicle — Miller says he would have stopped Arc a year earlier (Josh Miller / The Browser Company).

Risks and what didn't work

Novelty tax and complexity. "For most people, Arc was simply too different, with too many new things to learn, for too little reward" (Josh Miller / The Browser Company).
Not open-sourcing. Fans asked; Miller says Arc depends on proprietary ADK, also core to Dia, so open-sourcing Arc would give away Dia's IP (Josh Miller / The Browser Company, TechCrunch).
User trust hit. Sunset after a pivot and acquisition left many users disappointed on social media, per TechCrunch (TechCrunch).
Engagement percentages are company-reported in Miller's letter.

FAQ

Short answers drawn from the sections above.

What is Arc?

Why did Arc work?

Taste and craft. Early adopters loved the Nintendo/Disney-like product feel Miller described (Josh Miller / The Browser Company).
Strong D1 for a niche. Miller writes day-one retention was strong and stayers were "fanatics," with metrics more like a specialized pro tool than mass consumer software (Josh Miller / The Browser Company).

What can indie builders learn from Arc?

Measure revealed preference. Miller cites that only 5.52% of DAUs use more than one Space regularly, 4.17% use Live Folders, and 0.4% use Calendar Preview on Hover — vs Dia features used by ~40% and ~37% of DAUs (Josh Miller / The Browser Company).
Switching costs for browsers are high; novelty without immediate value is a tax, not a moat (Josh Miller / The Browser Company).
If the mission is mass-market infrastructure, a beloved niche product can still be the wrong vehicle — Miller says he would have stopped Arc a year earlier (Josh Miller / The Browser Company).

What didn't work for Arc, and what are the risks?

Novelty tax and complexity. "For most people, Arc was simply too different, with too many new things to learn, for too little reward" (Josh Miller / The Browser Company).
Not open-sourcing. Fans asked; Miller says Arc depends on proprietary ADK, also core to Dia, so open-sourcing Arc would give away Dia's IP (Josh Miller / The Browser Company, TechCrunch).
User trust hit. Sunset after a pivot and acquisition left many users disappointed on social media, per TechCrunch (TechCrunch).
Engagement percentages are company-reported in Miller's letter.

Sources

Every link cited above, in order of first use.

1.Josh Miller / The Browser Company (browsercompany.substack.com)
2.TechCrunch (techcrunch.com)

Arc: beloved browser, novelty tax, then a hard pivot to Dia and a sunset