Ladybird: a browser engine from scratch

You know that feeling when every browser is either Chromium-based or Firefox, and both are controlled by companies with ad-driven incentives? Mozilla just changed its terms of service regarding user data, Google controls both the dominant browser engine and the web standards body, and Microsoft abandoned its own engine for Chromium. There is no independent browser engine in active development — until Ladybird.

Think of Ladybird like building a car engine from raw metal instead of buying one from Toyota. Each browser tab runs in its own sandboxed process (separate renderer per tab), and networking and image decoding happen in their own isolated processes too. The JS engine has no JIT compiler — it interprets bytecode directly, which is simpler and safer. A new component called AsmInt uses hand-written assembly for hot paths (arithmetic, comparisons, property access), compiling via a custom DSL to native x86_64 and AArch64. The engine is written in C++ with a growing Rust component — the JS parser and regex engine are already Rust.

• No JIT compiler for JS/Wasm — interpreter-only approach reduces attack surface and implementation complexity, with AsmInt hand-written assembly partially closing the performance gap (1.37x-1.69x gains in March 2026)
• Multi-process sandboxed architecture — each tab gets its own renderer process, isolated from the system; image decoding and networking run out-of-process for robustness against malicious content
• Rust adoption accelerating — JS parser, regex engine, and bytecode generator ported to Rust; off-thread JS parsing enabled because 'since the parser is written in safe Rust, parallelizing it was straightforward'
• 2 million+ Web Platform Tests passed — crossed the 2M WPT mark in March 2026, up 5,292 subtests in a single month; GMail, Google Calendar, Figma, and YouTube (4K quality selection) all load
• Independent non-profit governance — 501(c)(3) with no search deals, no user monetization, no board seats for sale; funded by donations and sponsorships with an 18-month runway target
• Cross-platform build — runs on Linux (Qt6), macOS (AppKit), Windows (WSL2), FreeBSD; builds with CMake + Ninja + vcpkg, requires C++23 compiler and Rust toolchain

Who It Is For

If you're a C++ or Rust systems programmer interested in browser internals, web standards, or building engines from scratch. Also relevant if you care about web platform independence and want to follow (or contribute to) the only new browser engine in active development. Not useful if you need a daily-driver browser today — it's pre-alpha and explicitly 'only suitable for use by developers.'

Worth Exploring?

Clone it, build it, and test your own websites against it — the build is straightforward (`./Meta/ladybird.py run`) and the codebase is well-structured for learning. It's experimental and pre-alpha, not production-ready, but the project velocity is real: 352 PRs merged in March 2026 from 49 contributors. The Alpha release is targeted for 2026 on Linux and macOS. Worth following if browser engine diversity matters to you.