Getting Started with RDesk

# Verify RDesk is installed correctly
packageVersion("RDesk")
#> [1] '1.0.4'
# Check that the example app is bundled (if existing)
app_path <- system.file("templates/hello", package = "RDesk")
if (nzchar(app_path)) {
  file.exists(app_path)
}
#> [1] TRUE

The “Last Mile” of R Deployment

Data analysts build powerful tools in R, but sharing them with non-technical users is difficult. Shiny is excellent for web dashboards, but it requires a server and a steady network connection.

RDesk solves this by turning your R code into a native Windows desktop application. It runs as a standalone .exe with zero network overhead and no server required.

1. Installation

RDesk works out-of-the-box on modern Windows systems. You only need R and Rtools installed.

# Install from GitHub
# devtools::install_github("Janakiraman-311/RDesk")
library(RDesk)

2. Your First App (The One-Click Way)

The fastest way to build an RDesk application is using the built-in scaffolding tool. This generates a professional dashboard with a sidebar, KPI cards, and an asynchronous charting engine.

library(RDesk)

# Create a professional dashboard in a new directory
RDesk::rdesk_create_app("MyDashboard")

RDesk Hero Dashboard

Navigate to the MyDashboard folder and run app.R to see the results.

3. Building a Standalone Executable

The core value of RDesk is the ability to turn your R code into a standalone .exe that runs on any machine without requiring a pre-installed R version.

First build note: The initial build takes 3-5 minutes as RDesk downloads and bundles a portable R runtime. Subsequent builds complete in under 30 seconds.

# Build a portable ZIP bundle
RDesk::build_app(
  app_dir  = "MyDashboard",
  app_name = "MyProfessionalApp"
)

The resulting .zip file in the dist/ folder contains exactly what you need to distribute your tool. The end-user simply unzips and runs MyProfessionalApp.exe.

Using renv for reproducibility

If your project uses renv, commit your renv.lock in the project root. RDesk’s GitHub Actions workflows will automatically restore from that lockfile before building. Additionally, build_app() writes a bundle-level renv.lock into each distributable when renv is active.

4. Key Concepts

Zero-Port IPC

Zero ports, zero firewall issues. Communication happens over standard input/output using a high-speed JSON protocol.

Async Processing

Use the async() wrapper to run heavy R tasks without freezing the UI:

app$on_message("heavy_task", async(function(msg) {
  Sys.sleep(2) # Simulate work
  list(status = "Done")
}, app = app))