Deno 2.9: Native Desktop Apps and Beyond

Note: Check out Claw Patrol, the new open-source security firewall designed specifically for AI agents!

Released on June 25, 2026, by Bartek Iwańczuk, Deno 2.9 introduces a paradigm shift in how developers build desktop software. The headline feature is deno desktop, which allows you to create native applications using your existing web stack—~~without the heavy Electron boilerplate~~—resulting in a single, streamlined binary.

🚀 Getting Started

To move to the latest version, simply execute: deno upgrade

If you haven't installed Deno yet, choose your platform:

macOS/Linux (Shell): curl -fsSL https://deno.land/install.sh | sh
Windows (PowerShell): iwr https://deno.land/install.ps1 -useb | iex

🖥️ Introducing `deno desktop`

Traditionally, building desktop apps required complex toolchains like Tauri or Electron. deno desktop simplifies this by turning a script or web project into a self-contained application. The UI is rendered in a webview, while the backend logic is powered by Deno.

How it Works

Quick Implementation

Using Deno.serve() inside a desktop entrypoint automatically handles port binding, removing the need for manual wiring.

// main.ts
Deno.serve(( ) => new Response("<!DOCTYPE html> <h1>Hello from Deno desktop 👋</h1>", { 
  headers: { "content-type": "text/html" } 
}));

Run it with: $ deno desktop main.ts

Framework Integration

Deno 2.9 features automatic detection for several popular frameworks. If you run deno desktop . in a project folder, it will build and wrap the following:

Next.js / Remix / Nuxt
Astro / Fresh / SvelteKit
SolidStart / TanStack Start
Vite SSR

Pro tip: Use the --hmr flag during development for Hot Module Replacement.

Native Desktop APIs

You can now access OS-level features via the Deno.* namespace without external dependencies:

Deno.BrowserWindow: Manage window dimensions, visibility, and menus. Use window.bind() to create a bridge between the webview's JavaScript and the Deno backend.
Deno.Tray & Deno.Dock: Create system tray icons and macOS dock integrations.
Native Dialogs: Standard functions like alert(), confirm(), and prompt() now trigger native OS dialogs.
Deno.autoUpdate(): A built-in polling system for background binary patching.

Example: System Tray

// tray.ts
const tray = new Deno.attachPanel({ 
  url: "https://localhost:8000/panel" 
});

Choosing Your Backend

You can specify the rendering engine using the --backend flag:

Backend	Engine Used	Pros	Cons
`webview` (Default)	OS Native (WebView2/WebKit)	Small binaries, fast launch	Rendering varies by OS
`cef`	Bundled Chromium	Identical rendering everywhere	Larger binary, slower build

Commands:

$ deno desktop main.ts (Native)
$ deno desktop --backend cef main.ts (Chromium)

📦 Distribution & Cross-Compilation

Since deno desktop leverages the deno compile engine, it produces standalone binaries.

Supported Formats:

macOS: .app, .dmg
Windows: .exe, .msi
Linux: .AppImage, .deb, .rpm

You can build for multiple platforms from a single machine (e.g., a Linux CI runner) using the --target or --all-targets flags. The .msi and .deb/.rpm installers are written in Rust, meaning no platform-specific toolchain is required on the host.

# Build for current host
$ deno desktop --output MyApp.dmg main.ts 

# Cross-compile to Windows
$ deno desktop --target x86_64-pc-windows-msvc main.ts 

# Build for all 5 targets (Linux x64/arm64, Win x64, Mac x64/arm64)
$ deno desktop --all-targets main.ts

Note: Use --compress to create a self-extracting bundle for smaller file sizes.

Real-world Example: Check out denidian, a note-taking app built entirely with this workflow.

⚙️ Other Major Updates

Seamless Node.js Migration

Switching to Deno is now a matter of a few commands rather than a full migration. deno install now natively parses lockfiles from:

npm
pnpm
yarn
Bun

General Improvements

CSS: Now supports module imports.
Testing: A significantly more robust test runner.
Compatibility: Full support for Node.js 26.

📈 Performance Benchmarks

Deno 2.9 delivers massive improvements in memory efficiency, HTTP throughput, and startup speed.

Cold Start Comparison

The improvement in startup time is calculated as: $\text{Speedup} = \frac{\text{v2.8 Start Time}}{\text{v2.9 Start Time}} = \frac{34.2\text{ms}}{17.3\text{ms}} \approx 1.98\times$

`Deno.serve` Throughput

Measured on a dedicated x86_64 Linux machine with a concurrency of 100:

Metric	Deno 2.8	Deno 2.9	Result
Cold Start	$34.2\text{ms}$	$17.3\text{ms}$	$\downarrow$ Faster
Real-world Req/s	$56.8\text{k}$	$72.4\text{k}$	$\uparrow$ Higher