One website, every screen size at once: our own Mac browser, from first day of work to a released beta in 12 days

The challenge

Every team that builds a website faces the same chore: checking that pages look right on desktop, tablet, and phone. Today that means dragging a browser window narrower and wider, switching between simulated devices one at a time, or paying for a service that takes screenshots on remote computers. None of these show all the target sizes live and side by side, and none handle a site that is still in development on your own machine. The check-and-recheck loop costs time on every design change.

There is a second user the existing tools ignore entirely: AI agents. AI agents increasingly write website code, but they have no reliable way to check their own work across screen sizes. The built-in tools are made for a person at a keyboard, not for software that needs to open a page, click through it, and collect screenshots on its own.

We built VarioBrowser to close both gaps: one glance for people, one set of commands for AI agents.

What we built

VarioBrowser is a browser for the Mac. You open one website and the app shows it at desktop, tablet, and phone sizes at the same time - plus any custom sizes - as real, live views on a single zoomable canvas. Sign in once and every screen size is signed in.

The views stay in sync. Scroll one and the others follow. Click a button in one view and the same button is clicked in the rest, even though each size lays the page out differently. Type into a form and the text appears everywhere, with safeguards so actions never echo back and forth or trigger duplicate submissions.

Alongside the app sits an interface built for AI agents: around 30 commands cover opening pages, clicking, typing, taking screenshots, and pulling diagnostic detail about what the page is doing. The app installs this integration itself, so an AI agent on the same Mac can test a website at every screen size without anyone setting anything up.

The paid tier adds simulation of common vision deficiencies, including in screenshots, an instant-replay recorder that keeps the last moments on screen so you can save a video of a bug after it has already happened, and bug-context capture that gathers what the page was doing - with login tokens and other secrets stripped out automatically before anything is saved.

Around the app we built an 8-page marketing website, a licensing system that handles accounts and paid tiers, and the release work to ship one product two ways: approved for the Mac App Store and also available as a direct download from our site.

How we built it

Three decisions shaped the build. First, tight scope: we cut everything that did not serve the core promise, which is how the work fit into 12 days. Second, AI agents as first-class users: the agent interface is not an add-on bolted onto the app - everything the app can do, an agent can do too, which is what makes the agent use case real rather than aspirational. Third, shipping all the way: store approval, packaging, accounts, and licensing - the unglamorous last mile that separates a demo from something people can install and pay for.

What this build demonstrates

VarioBrowser is our own product, built the way we build for clients. Twelve days cover the full distance: a working Mac app, an automation surface for AI agents, a marketing website, a licensing system, and a release through the Mac App Store plus a direct download. It is currently available as a time-limited beta. The same approach - small scope, AI agents as real users, shipping discipline - is what we bring to client engagements.