RDesk

Package R analyses into self-contained native desktop applications with zero network port requirements.

---

RDesk packages your R analysis into a standalone desktop application for Windows, macOS, and Linux. Instead of running an HTTP server and opening a browser tab, RDesk uses a native launcher plus an embedded webview so the app runs offline with local IPC only.

Quick Start

# Install from CRAN
install.packages("RDesk")

# Scaffold an interactive app
RDesk::rdesk_create_app("MyApp")

# Run it immediately
source("MyApp/app.R")

When you are ready to ship:

RDesk::build_app(
  app_dir         = "MyApp",
  app_name        = "MyApp",
  build_installer = TRUE
)

# Windows -> dist/MyApp-1.0.0-setup.exe
# macOS   -> dist/MyApp.app and dist/MyApp-1.0.0.dmg
# Linux   -> dist/MyApp-1.0.0/ and dist/MyApp-1.0.0.tar.gz

Windows can produce an installer, macOS can produce an .app bundle and optional .dmg, and Linux produces a portable bundle plus .tar.gz. No separate R installation or browser server is required on the target machine when using the bundled runtime.

Development version

devtools::install_github("Janakiraman-311/RDesk")

Why RDesk?

Feature	Shiny	RInno (archived)	Electron + R	RDesk
Network ports required	Yes	Yes	Yes	No
Works fully offline	No	Partial	Partial	Yes
Native file dialogs and menus	No	No	Via JS	Yes
Distribution format	Server URL	Installer	Installer	Bundle or installer
Bundle size	Server-side	350 MB+	350-500 MB	~200 MB
Skills required	R	R + Electron	R + Node.js	R only

Core Features

• Zero-port IPC - R and the UI communicate through native stdin/stdout pipes and platform webview bindings, without opening network ports.
• Async by default - Non-blocking background work using mirai and callr with support for progress updates, loading states, and cancellation.
• Version-safe runtime - build_app() bundles a matching R runtime so the shipped app and its packages stay aligned.
• Modern web UI - Build the interface with plain HTML, CSS, and JavaScript while keeping the backend in R.
• Automated scaffolding - rdesk_create_app() generates a working application template with handlers, structure, and theme support.
• Native packaging - Produce platform-specific bundles and installers without adopting a browser-server deployment model.

Who It’s For

RDesk is built for R developers who need to package a local statistical or data tool for non-R users.

• Pharma and clinical - distribute review or validation tools for offline execution.
• Consulting - package analytical models into branded tools without exposing source code.
• Internal teams - replace complex spreadsheet macros with structured R applications.
• Restricted environments - ship local tools where listening ports are disallowed or heavily scrutinized.

Example Apps

Two apps ship with RDesk demonstrating different complexity levels.

CarsAnalyser - minimal dashboard

app_dir <- system.file("apps/mtcars_dashboard", package = "RDesk")
source(file.path(app_dir, "app.R"))

Data Intelligence Studio - full-featured data profiling tool

app_dir <- system.file("apps/data_studio", package = "RDesk")
source(file.path(app_dir, "app.R"))

Documentation

Full documentation is available at janakiraman-311.github.io/RDesk.

Guide	What it covers
Getting Started	From install.packages() to your first native app
Coming from Shiny	Side-by-side mapping of common Shiny patterns to RDesk
Async Guide	Background tasks, progress overlays, cancellation
Cookbook	Practical desktop-app recipes
Why RDesk?	Project background and architecture

License

MIT (c) Janakiraman G.