Performance

written for Apple silicon, not ported to it.

A native shell around a tuned WebKit. Every number below comes from a MacBook Air M4, 16GB, warm cache, sixty tabs open.

0.4s
cold launch to first tab
1.4×
faster page loads
1.9 GB
memory, sixty tabs open
18h
battery on a MacBook Air
01

benchmarks.

Same machine, same network, same sixty tabs.

Memory footprint

Terrava
1.9 GB
Browser C
3.1 GB
Browser F
3.4 GB
Browser E
2.6 GB

Median page load

Terrava
380ms
Browser C
540ms
Browser F
610ms
Browser E
470ms

MacBook Air M4, 16GB, warm cache, sixty tabs open, on macOS 15.4. The comparison browsers stay unnamed until the method and the raw runs are published.

02

how.

Three decisions doing most of the work.

Predictive wake

Terrava starts restoring a sleeping tab the moment your pointer heads for it, so the click lands on a live page.

One process per space

Isolation where it matters, shared memory where it does not. Sixty tabs do not mean sixty processes.

Energy-aware timers

Background scripts are coalesced onto efficiency cores. The fans stay off and the battery indicator stays boring.

fast enough to forget about.

Free, 96 MB, universal binary. Import your tabs and see for yourself.