Conversations
Rust GUI ecosystem heats up driven by AI, but routes remain fragmented
Raph Levien wrote an article reviewing the development of Rust GUI toolkits, saying that four and a half years after his previous recommendations article, about 20 Rust GUI toolkits have been released, and AI is increasing developer interest in Rust interfaces. The article argues that Rust remains attractive for performance, safety, and memory footprint, but there is no consensus on what approach to use for the UI layer, and native choices for desktop and mobile platforms each have their own complexities.
The article notes that there are still strong reasons to build graphical interfaces with Rust. Many cross-platform interface applications have performance problems, platform-bundled GUI toolkits have stagnated, and system vendors have not actively solved the cross-platform unification problem, which is why Electron became popular. But memory costs and the shift in focus toward mobile make bloated size carry a real price, making high-performance Rust UI a viable alternative.
Adoption of Rust UI has not been fast over the past few years, and there are few flagship applications. The author mentioned that the Zed editor, based on its own GPUI framework, is an important example that Rust interfaces can actually ship; GPUI has also drawn attention because of AI-driven rewrite projects such as Discord. However, the author believes GPUI is better suited to scenarios like Zed, and its 2D rendering capabilities fit text editors, but it is not yet equivalent to renderers such as Vello or Skia in gradients, compositing, filters, and other areas.
AI is changing adoption barriers. Traditional UIs commonly use TypeScript, Kotlin, and Swift, while Rust has higher learning and hiring barriers; but the article says there are signs that the cost of large language models writing Rust is not very different from other languages, while Rust’s type system may make large codebases easier to maintain. Based on this, the author judges that Rust is rapidly becoming a choice for AI rewrites of existing codebases, but UI-layer strategies will still see many different attempts.
The article also discusses the concept of “native.” Windows still lacks a clear and consistent native GUI choice, while on macOS there is a split between AppKit and SwiftUI, and SwiftUI also has performance issues. Mobile is different: platforms continue to invest, and applications need native look and feel in details, such as iOS’s Liquid Glass. The material cuts off there and does not present subsequent conclusions.