vhsR

Animated GIF demos of R code, generated by an R one-liner. Wraps charmbracelet/vhs.

Installation

# install.packages("pak")
pak::pak("schochastics/vhsR")

vhs itself plus its runtime dependencies (ttyd, ffmpeg) are downloaded into a per-user cache on first use:

library(vhsR)
vhsr_install()   # one-time setup
vhsr_doctor()    # confirms everything is in place

On macOS, vhsr_install() will print a brew install line for ttyd and ffmpeg (those have no usable upstream binaries for macOS); on Linux and Windows everything is downloaded automatically.

Usage

record_demo({
  x <- 1:5
  mean(x)
  head(cars, 3)
}, output = "demo.gif")

That writes demo.gif next to your working directory. The GIF shows the code being typed at a real R prompt, with output appearing as it would in an interactive session.

Styling

record_demo(
  {
    fit <- lm(mpg ~ wt, data = mtcars)
    summary(fit)$coefficients
  },
  output         = "fit.gif",
  width          = 1000,
  height         = 500,
  font_size      = 20,
  theme          = "Dracula",
  typing_speed   = "100ms",
  playback_speed = 1.25
)

Pacing

Four extra timing args control how the recording feels:

• start_pause — sleep between R startup and the first typed line (default "800ms").
• end_pause — sleep before the (hidden) q() exit (default "500ms").
• paragraph_pause — when set, blank lines in expr become a sleep of this duration instead of being typed.
• typing_speed_jitter — numeric in [0, 1]; per-line typing speed is sampled within ±jitter of typing_speed so the recording feels less mechanical. Call set.seed() first for reproducibility.

record_demo(
  {
    x <- 1:5
    mean(x)

    head(cars, 3)
  },
  output              = "paced.gif",
  paragraph_pause     = "1s",
  typing_speed_jitter = 0.3
)

Recording from a script file

If the demo lives in a .R file, point at it directly:

record_demo_file("examples/demo.R", output = "demo.gif")

Equivalent to inlining the file’s contents in record_demo(). Syntax errors are caught up-front with a pointer to the file path.

Single-frame screenshot

For static documentation — prompt, code, and output in a PNG instead of a GIF:

record_demo_screenshot(
  {
    fit <- lm(mpg ~ wt, data = mtcars)
    summary(fit)$coefficients
  },
  output = "fit.png"
)

By default the screenshot is taken after the last typed line. Pass at = "after:N" (1-indexed) to capture after a specific line.

Knitr / Rmd chunks

vhsR registers a vhsr knitr engine on package load, so you can drop a chunk into a vignette, README.Rmd, or pkgdown article:

``` vhsr
x <- 1:5
mean(x)
head(cars, 3)
#> ![](man/figures/demo.gif)
```

The chunk records a GIF at output= and inserts the corresponding markdown image-include in the rendered document. Forwarded chunk options: output, width, height, font_size, theme, typing_speed, playback_speed, line_pause, backend.

Embedding in HTML

For Rmarkdown documents or Shiny apps that want playback controls:

vhsr_widget("demo.mp4")

Returns an htmltools::tagList. .gif becomes a styled <img>; .mp4 / .webm become a <video controls loop muted playsinline> with the appropriate MIME source. width / height accept either numerics (treated as pixels) or strings ("100%", "50vw", …).

Escape hatch

For full control over the recording, write a .tape script directly and hand it to vhsr_run_tape() — or pass the tape = argument to record_demo():

vhsr_run_tape('
Output hello.gif
Set FontSize 22
Type "echo hi from vhs"
Enter
Sleep 1s
')