We didn't translate a single word. Developers from 90+ countries found us anyway.

Last week we open-sourced Shuru — a microVM sandbox that boots disposable Linux environments on macOS. We posted it on Hacker News, went back to work, and by evening it was #1 on Show HN.

What happened next we didn't plan for.

The numbers

Within a few days: 211 HN upvotes, 80 comments (zero negative), 469 GitHub stars, and visitors from over 90 countries. All organic. No ads, no paid outreach, no influencer partnerships.

Here's the part that caught us off guard: the international breakdown.

The US was expected. 33% of traffic, solid engagement. That's where HN's audience skews.

But then:

• Germany (7.7%) and UK (6.7%) showed up strong – Europe's developer community picked it up immediately

• South Korea (2.4%) surprised us the most. A site called news.hada.io, basically the Korean Hacker News, surfaced Shuru organically. 185 visitors, 51 seconds average time on site. Nobody asked them to post it.

• India (4.8%) had the highest engagement of any major country: 67 seconds average session time, 83% bounce rate. For comparison, the US averaged 44 seconds. Indian developers weren't just clicking — they were reading.

• Japan, Singapore, Vietnam, Indonesia – all sent meaningful traffic with real engagement.

We had visitors from countries we'd never have thought to target: Senegal, Armenia, Kosovo, Paraguay, Faroe Islands.

What we didn't do

We didn't localize anything. The site is English-only. The README is English-only. We have no social media presence in any of these markets. We didn't post in any Korean, Japanese, or German developer community.

So why did it spread?

What actually happened

Developer tools are a universal language. If your tool solves a real problem and the interface is a terminal, there's no localization barrier. shuru run reads the same everywhere. The code examples work the same everywhere.

Aggregator networks are global. When you hit the HN front page, it doesn't just stay on HN. There's a constellation of sites that mirror or curate HN content – hada.io (Korea), hckrnews.com, hackerweb.app, brutalist.report, hn.algolia.com, and dozens of others. These sites have their own communities that discover and engage independently. We tracked traffic from 15+ HN-adjacent sites.

Good defaults travel. The thing people responded to wasn't a feature — it was the philosophy. "Everything is off unless you turn it on." No networking by default. No shared directories. No persistence. That idea doesn't need translation. Developers everywhere have the same frustration with tools that are configured for the wrong defaults.

The landing page did the heavy lifting. shuru.run showed exactly what we built – feature cards, interactive terminal demos, use cases, and a two-line install. Someone could understand what Shuru does and why it exists in 30 seconds without ever opening the repo. That's the universal onboarding.

What we learned

Don't target countries. Target problems. We never thought about "how do we reach Korean developers". We thought about "how do we make sandboxing dead simple". The Korean developers found us because the problem is universal.

Product Hunt sent us 14 visitors. All bounced. Our PH launch was the same week. The audience there wants visual, demo-friendly products. A CLI sandbox is not that. Meanwhile, HN – where people understand what Virtualization.framework means – gave us 4,000+ visitors with 80 positive comments. Know where your people are.

Aggregator effects are real. A single HN post became distribution across 15+ sites, multiple languages, and 90+ countries. You can't manufacture this – but you can make it more likely by building something that solves a specific, real problem and making it dead simple to understand from a README.

The backstory

I built Shuru because I needed it. I was working on Neko, a lightweight AI agent runtime, and the agent needed to execute arbitrary code safely. Docker shares the host kernel, so it is not real isolation. Lima needed too much config. So I built a tool that boots real VMs using Apple's Virtualization.framework, with safe defaults, in one command.

Turns out a lot of other developers, in a lot of countries, had the same problem.

Site: shuru.run