v0.1.2082 Sep 2026·JavaScript & JIT, Web APIs & DOM, Memory & reading

Floati's own engine got dramatically faster at the code real pages run every frame

The from-scratch engine has a fast lane - a just-in-time compiler, which watches a page's JavaScript as it runs and turns the parts that repeat straight into native machine code, the fastest form a processor can execute. Until now that fast lane covered arithmetic, walking arrays by index and a handful of other shapes, but bailed back to the slow interpreter for several things pages do constantly - building an array, building an object, walking a list with for-of, joining strings together, and branching on a value's type. All of those now compile to native code. Separately, Map and Set - two of the most-used data structures in modern JavaScript, the in-memory tables real apps keep their working data in - used to get slower the more you put in them; they are now instant to read and write no matter how large they grow. None of this changes what a page shows or does: every fast path is proven, byte for byte, to give the exact same answer as the slower one it replaces. It is one more stretch of the road toward Floati's engine rendering everything you browse on its own, fast enough that you never notice which engine drew the page.

Parity with the reference browser

74overall, as of v0.1.343

Web APIs / DOM surface42

Security architecture66

GPU / graphics acceleration68

DevTools / extensions72

Networking85

Media85

Accessibility85

Real-world site compatibility99

HTML parsing100

CSS parsing + selectors100

Layout: block / flex / grid / tables / writing-modes100

Paint + visual100

JavaScript language100

JavaScript performance (JIT)100

HTML parsing100

What changed 3 changes

JavaScript & JIT	Building arrays and objects, and looping over them with for-of, now compile to native machine code. Turning freshly-loaded data into a list of records - about the most common thing a real page's code does once it has loaded - used to fall back to the slow interpreter the moment it pushed to an array or built an object inside a loop; it now runs as compiled native code, several times faster, and the result is guaranteed identical to before.
JavaScript & JIT	Joining strings together in a loop, and branching on a value's type - the everyday `typeof x` checks and string comparisons a page uses to decide what to do next - now compile to native code as well. These are two more shapes pages run constantly that previously forced the slow path, so dispatch-heavy and text-building code speeds up with no change to its output.
Memory & reading	Map and Set, and their weak-keyed siblings, are now instant to read and write however large they grow. Before, every single lookup scanned the whole collection, so a Map holding ten thousand things was thousands of times slower than one holding ten - and code that leans on them could stall for seconds. The cost is now flat, the way these collections behave in every other browser. Insertion order, exact key matching and every observable behaviour are unchanged - only the speed.