Servo 0.4 Hits Record 558 Commits: Real Sites Now Work

Servo 0.4.0 shipped yesterday — July 31, 2026 — with 558 commits in a single month, a project all-time record. For the first time, the Rust-native browser engine renders Google Photos, Zulip, and lichess.org correctly, crossing a threshold that separates “experimental library” from “engine that can handle the real web.” Chromium controls 78.4% of global browser sessions. Servo is the only serious Rust-native alternative, and this release is proof it is no longer a thought experiment.

Servo 0.4: 558 Commits, Real Sites Now Working

Three months ago, Servo 0.1.0 shipped to crates.io — a milestone for embeddability, but the engine still stumbled on real-world pages. 0.4.0 is a different category of release. In June alone, the team landed 558 commits, up from roughly 130 per month in September 2023. That is a 4x velocity increase driven by Igalia’s engineering team, which now works on Servo full-time under Linux Foundation Europe governance.

The official June in Servo post confirms what the numbers suggest: “Layout correctness has significantly improved on lichess.org.” Zulip and Speedtest.net work thanks to improved variable font handling. Google Photos renders acceptably. Cash Converters is functional. Google Maps and OpenStreetMap display but still have interactivity issues — a known gap, not a surprise. This is a meaningful step, not a PR spin.

What Actually Shipped in This Release

Beyond site compatibility, the release adds SharedWorker API support (available under --pref dom_sharedworker_enabled), with ServiceWorker also experimental. CSS gains orientation, pointer, hover, device-width, and device-height media query features — the kind of queries that real sites actually depend on. The DOM gains console.dir(), textStream() methods, and improved pointer and touch event handlers.

The performance numbers matter for the embedded use case: BoxFragment memory shrank from 288 to 240 bytes — a 17% reduction. Canvas operations now use up to 23% less power. Sixteen crash bugs were fixed through fuzzing, addressing logic-level issues that Rust’s memory safety does not automatically prevent. These are not headline features, but they are exactly what you need before embedding a browser engine in production software.

Related: Rust Hits TIOBE Top 10 for the First Time: What It Means

The Problem Servo Is Actually Solving

Blink runs 78.4% of global browser sessions. Chrome alone is approximately 73%, and when you add Edge, Brave, Opera, and Vivaldi, Chromium powers over 80% of the web. Browser market share data for 2026 shows Gecko (Firefox) at 2.6% and falling — down 1.4 points over four years. We have quietly returned to the IE6 era, where one engine’s bugs and one company’s standards decisions affect nearly everyone.

A single Blink CVE now reaches 80% of web users. Google shapes web standards unilaterally through Chromium. Servo is not positioned to displace Chrome for consumers anytime soon — that would be overconfident. However, the embedded market is a different story. Servo ships as cargo add servo on crates.io. The oxibrowser project already uses it to power a CDP-compatible headless browser designed specifically for AI agents. Kiosk displays, desktop app WebViews, automated testing pipelines — these are the places where “not Chromium” is a practical choice today.

Key Takeaways

• Servo 0.4.0 is the first release that renders major real-world sites — Google Photos, Zulip, lichess.org — narrowing the gap between experimental and usable.
• 558 commits in June is an all-time monthly record; the 4x velocity increase since 2023 reflects Igalia’s full-time engineering commitment under Linux Foundation Europe.
• The near-term story is embedded and headless use cases, not consumer browsing — cargo add servo is live on crates.io, and AI agent browsers like oxibrowser are already building on it.
• Blink’s 78.4% browser session share is a structural risk to the open web; Servo is one of two serious non-Chromium alternatives, and it is the one written in Rust.