JavaScript runtime developed with AI: Edge.js for secure Node.js applications

The company Wasmer, which is behind the WebAssembly runtime of the same name, has released Edge.js. The open-source JavaScript runtime specializes in securely executing Node.js workloads in WebAssembly sandboxes. Edge.js is said to be particularly suitable for AI and edge computing use cases.

The new JavaScript runtime Edge.js is not to be confused with the older .NET project Edge.js.

Heise Conference: enterJS Integrate AI

Enhance web applications with AI so they become truly better? The online special event enterJS Integrate AI on April 28, 2026, shows how. Early bird tickets and group discounts are available in the online ticket shop.

Edge.js for secure Node.js applications

As the Wasmer team explains the reasons behind the new runtime, Node.js has two difficulties: it is tied to V8 as the only JavaScript engine and cannot securely execute workloads without containerization or hardware virtualization. Other providers like Deno or Bun have also recognized this and developed their own JavaScript runtimes, but according to Wasmer, Edge.js is the first fully sandboxed variant without Docker containers.

Edge.js uses the Node API (formerly N-API) as an abstraction layer, an API for creating native add-ons that is maintained as part of Node.js and abstracts away the V8 engine. This allows JavaScriptCore and QuickJS to be used as JavaScript engines in Edge.js. WASIX is used for system call sandboxing, a superset of the modular system interface for WebAssembly called WASI (WebAssembly System Interface).

According to Wasmer, Edge.js relies on compatibility with Node.js 24 and can execute everything that Node.js can execute, including all corresponding frameworks like Next.js or Astro. Currently, Edge.js 5 is five to 20 percent slower than Node.js in native execution and 30 percent slower with full sandboxing using Wasmer. Application startup times are also slower than with Node.js. The development team plans to work on speed on the way to Edge.js 1.0. A concrete goal is for Edge.js to be faster than Bun or Deno for most workloads.

Via AI coding to Edge.js

The Wasmer team states that they used artificial intelligence, mainly OpenAI GPT-5.4, for the implementation of Edge.js. A smaller startup like Wasmer would otherwise have needed at least one or two years for this project instead of just a few weeks. Thanks to the AI coding agent OpenAI Codex, developers in the team without expertise in C++ or Node.js could also contribute to bug fixing.

Further information on the initial Edge.js release is available on the Wasmer blog.

(mai)