Distill the web into a native binary.

Attar versus Electron, on the same app.

balena Etcher as Electron ships it, and balena Etcher as Attar builds it from the same source. Medians; lower is better.

apps

balena Etcher on Electron 37.2.4 and the same app compiled by Attar

machine

Apple M5 Max, macOS 27, one machine for both sides

method

5 launches timed to the first window through the Accessibility API; 3 idle samples of 60 s each; footprint and CPU summed over every process of the app

caveat

taken beside four concurrent builds (load 5 to 12), so a comparison, not a ceiling; on a quiet machine Attar reaches its first frame in 274 ms

What is inside the binary.

The same app, to one scale. Hover a part.

Own the giant. Carve away its bulk.

A real Chromium fork. Blink stays and runs your app; the browser around it does not ship.

Anything the compiler cannot prove is a build error, not a silent guess.

Can I use it in Attar?

Explore support across Attar, Attar AOT and Attar Electron: HTML, CSS, JavaScript, React and desktop APIs. Each feature has a scope note and an assessment date.

Not a smaller binary. A binary you can reason about.

Size and speed are consequences. The product is knowing exactly what remains, why, and what it cannot do.

Write the stack you already know.

React-shaped components, HTML and CSS, typed TSX. Or an existing Electron app, as is.

export function Counter() {
  const [n, setN] = useState(0)
  return (
    <button class="tally" onClick={() => setN(n + 1)}>
      pressed {n} times
    </button>
  )
}

.tally {
  font: 600 15px/1 system-ui;
  padding: 10px 16px;
  border-radius: 8px;
  background: #ffb23e;
  color: #1a1206;
}
.tally:active {
  transform: scale(.97);
}

Boundaries are explicit. What Attar can't prove in a closed world is a compile error that names the reason.

Attar is the first product of Occam.

One method: keep the giant as the source of truth, move it out of runtime and into the build.