Built from nothing, catching up fast
Most browsers in use today trace back to one of two rendering engines: Chromium or WebKit. Ladybird is neither. It is an independent browser engine built from scratch, with its own HyperText Markup Language (HTML) parser, its own layout system, and its own JavaScript engine, and its August 2026 progress report shows the kind of jump that makes a from-scratch engine worth watching again.
On the Speedometer 2 benchmark, a standard measure of how a browser handles realistic web application workloads, Ladybird's score rose from about 47 to about 64 over the month. Speedometer 3, a newer and more demanding version of the same benchmark, went from roughly 2.5 to 3.9. The bigger jump came on StyleBench, a benchmark focused on cascading style sheets (CSS) processing speed, where Ladybird's score climbed from about 3.5 to about 83, a roughly twelve-fold improvement that puts it in range of the major production engines it is trying to catch up to.
Where the speed came from
The gains were not one lucky optimization. Ladybird's team rewrote its CSS parsing, computed-style storage and computation, and its painting and display-list pipeline, moving that code from C++ into Rust. Layout results are now cached and reused, so many style changes skip a full page relayout rather than recalculating everything from scratch. The document object model (DOM) was also restructured so that a DOM node no longer carries JavaScript object state until a script actually touches it, which cuts memory use for pages that do not need it.
Individual operations improved by wide margins: canvas fillStyle parsing went from 90 times slower than Chromium to about 2 times slower, the replaceChildren() DOM method went from 13 times slower to under 4 times slower, and list operations went from 120 times slower to under 20 times slower. None of those numbers put Ladybird ahead of Chromium yet, but the gap closing this fast in a single month is unusual for a project this size.
Catching up on media and features, not just speed
Ladybird also added support for fragmented Moving Picture Experts Group-4 (MP4) files in its Media Source Extensions implementation, along with the Advanced Video Coding (AVC), High Efficiency Video Coding (HEVC), AOMedia Video 1 (AV1), and Advanced Audio Coding (AAC) codecs. In practical terms, that is what let the team get Twitch streaming working and improved YouTube playback, both of which rely on those same underlying pieces. New CSS support landed for scroll-snap properties, and CSS animations, including filter animations, now run off the browser's main thread rather than competing with it.
The built-in developer tools gained JavaScript breakpoints, line-by-line stepping, and watch expressions, bringing Ladybird's debugging experience closer to what developers already expect from Chrome or Firefox.
An alpha that is waiting on infrastructure, not code
Ladybird's team has said its first public alpha is still targeted for 2026, and notably, the blocker at this point is not engine completeness but infrastructure: crash reporting, signed builds, automatic updates, and distribution channels all need to be in place before a wider release makes sense. That is a meaningfully different kind of "almost ready" than a browser that still needs core rendering work.
The report drew attention beyond Ladybird's own newsletter. A discussion thread on Hacker News reached 288 points and 84 comments, with commenters expressing surprise that Twitch streaming already works this early in the project's life. Not all of the discussion was purely positive: some raised concerns that Ladybird has closed external code contributions to a self-selected group of maintainers, despite still describing itself as open source, and a maintainer reportedly said pull requests are now screened for signs of AI-generated code, using the quality of a contributor's messages as one signal of human authorship. Skeptics in the thread also pointed out that the compatibility gaps that remain, the long tail of real-world websites that do something a spec does not quite cover, tend to be the hardest part of building a browser engine, harder than any single benchmark number captures.
What it means for developers
None of this changes what developers need to test against today. Ladybird is not yet a production target, and the realistic timeline for a stable, general-use release still runs past this year. What is worth tracking is the trajectory: a team building a rendering engine from zero closed a twelve-fold gap on a CSS benchmark in a single month and got a mainstream streaming site working in the same window. If that pace holds, Ladybird's public alpha, whenever its infrastructure catches up to its engine, will be a more credible third option for web testing than a from-scratch browser project has been in years. Developers who care about the web staying open to more than two rendering engines have a concrete reason to check back on Ladybird's monthly reports rather than writing the project off as a curiosity.
