resmush

resmush is an R package for optimizing local and online image files, individually or in entire directories, with the reSmush.it API. The API is free for personal use and does not require an API key. reSmush.it is also available through WordPress and other tools.

The reSmush.it API provides:

Optimization without an API key.
Support for PNG, JPEG, GIF, BMP and TIFF files.
A per-file size limit: files must be smaller than 5 MB.
Compression powered by several algorithms:
- pngquant: Removes unnecessary data from PNG files while preserving full alpha transparency.
- jpegoptim: Lossless optimization based on Huffman table optimization.
- OptiPNG: A PNG optimizer used by several online compression tools.

[!NOTE]

This package is stable and maintained on a best-effort basis. Maintenance currently prioritizes CRAN compatibility, bug fixes and regressions over new features.

Installation

Install resmush from CRAN with:

install.packages("resmush")

Check the documentation for the development version at https://dieghernan.github.io/resmush/dev/.

You can install the development version of resmush from GitHub with:

# install.packages("pak")
pak::pak("dieghernan/resmush")

Alternatively, install resmush from r-universe:

# Install resmush in R:
install.packages(
  "resmush",
  repos = c(
    "https://dieghernan.r-universe.dev",
    "https://cloud.r-project.org"
  )
)

Examples

Optimize and download an online JPEG image with resmush_url():

library(resmush)

url <- "https://dieghernan.github.io/resmush/img/jpg_example_original.jpg"

resmush_url(
  url,
  outfile = "man/figures/jpg_example_compress.jpg",
  overwrite = TRUE
)
#> ══ resmush summary ═════════════════════════════════════════════════════════════
#> ℹ Input: 1 URL, 178.7 Kb total.
#> ✔ Optimized 1 URL: size is now 45 Kb (was 178.7 Kb). Saved 133.7 Kb (74.82%).
#> Saved result in directory 'man/figures'.

(a)

(b)

Figure 1: Original image (a) (178.7 KB) and optimized image (b) (45 KB), a 74.8% reduction in file size. Click to enlarge.

Use the qlty argument to adjust the JPEG quality level. For best results, use values above 90.

# Use a low JPEG quality level.
resmush_url(
  url,
  outfile = "man/figures/jpg_example_compress_low.jpg",
  overwrite = TRUE,
  qlty = 3
)
#> ══ resmush summary ═════════════════════════════════════════════════════════════
#> ℹ Input: 1 URL, 178.7 Kb total.
#> ✔ Optimized 1 URL: size is now 2.2 Kb (was 178.7 Kb). Saved 176.4 Kb (98.74%).
#> Saved result in directory 'man/figures'.

Figure 2: Image with visible compression artifacts caused by high compression (qlty = 3).

All optimization functions invisibly return a data frame with one row per result and columns containing source and destination paths, formatted file sizes, file sizes in bytes, compression ratios and status notes. They return NULL if no result is available. Successful API calls also write the optimized files to disk. The following example shows the result for a local image file:

png_file <- system.file("extimg/example.png", package = "resmush")

# Copy to a temporary file for this example.
tmp_png <- tempfile(fileext = ".png")
file.copy(png_file, tmp_png, overwrite = TRUE)
#> [1] TRUE

summary <- resmush_file(tmp_png, overwrite = TRUE)

tibble::as_tibble(summary[, -c(1, 2)])
#> # A tibble: 1 × 6
#>   src_size dest_size compress_ratio notes src_bytes dest_bytes
#>   <chr>    <chr>     <chr>          <chr>     <dbl>      <dbl>
#> 1 239.9 Kb 76.1 Kb   68.29%         OK       245618      77896

Alternatives

Several other R packages provide image optimization tools:

The xfun package (Xie 2026b) provides:
- xfun::tinify(): Similar to resmush_file() but uses TinyPNG and requires an API key.
- xfun::optipng(): Compresses local files using OptiPNG. The program must be installed locally.
The tinieR package provides an R interface to TinyPNG.
The tinyimg package (Xie 2026a) optimizes local PNG and JPEG files using Rust libraries. It supports lossless PNG optimization via oxipng, optional lossy PNG palette reduction and JPEG re-encoding via mozjpeg.
The optout package is similar to xfun::optipng() but offers more options. It requires additional local software.

Tool	CRAN	Additional software	Online images	API key required	Limits
`xfun::tinify()`	Yes	No	Yes	Yes	500 compressions per month (free tier)
`xfun::optipng()`	Yes	Yes	No	No	None
tinieR	No	No	Yes	Yes	500 compressions per month (free tier)
tinyimg	Yes	Yes (Rust toolchain)	No	No	None
optout	No	Yes	No	No	None
resmush	Yes	No	Yes	No	Personal use only. Files smaller than 5 MB.

Table 1: Comparison of image optimization alternatives for R packages.

Tool	PNG	JPEG	GIF	BMP	TIFF	WebP	PDF
`xfun::tinify()`	✅	✅	❌	❌	❌	✅	❌
`xfun::optipng()`	✅	❌	❌	❌	❌	❌	❌
tinieR	✅	✅	❌	❌	❌	✅	❌
tinyimg	✅	✅	❌	❌	❌	❌	❌
optout	✅	✅	❌	❌	❌	❌	✅
resmush	✅	✅	✅	✅	✅	❌	❌

Table 2: Supported formats for R packages.

In practice, resmush is designed for quick image optimization with minimal setup, including support for online image files and formats such as GIF, BMP and TIFF. Packages such as tinyimg may be a better fit for fully local workflows focused on PNG and JPEG optimization and fine-grained control over compression settings.

Citation

Hernangómez D (2026). resmush: Optimize and Compress Image Files with reSmush.it. doi:10.32614/CRAN.package.resmush. https://dieghernan.github.io/resmush/.

A BibTeX entry for LaTeX users:

@Manual{R-resmush,
  title = {{resmush}: Optimize and Compress Image Files with {reSmush.it}},
  doi = {10.32614/CRAN.package.resmush},
  author = {Diego Hernangómez},
  year = {2026},
  version = {1.0.3},
  url = {https://dieghernan.github.io/resmush/},
  abstract = {Optimize and compress local and online image files with the reSmush.it API <https://resmush.it/api/>. Process individual files or entire directories. The API is free for personal use, accepts files smaller than 5 MB and supports PNG, JPEG, GIF, BMP and TIFF files.},
}

References

Xie, Yihui. 2026a. tinyimg: Optimize and Compress Images. https://doi.org/10.32614/CRAN.package.tinyimg.

Xie, Yihui. 2026b. xfun: Supporting Functions for Packages Maintained by Yihui Xie. https://github.com/yihui/xfun.