diffoscope

diffoscope finds out what makes two files or directories different. It recursively unpacks archives of many kinds and transforms binary formats into a human-readable form before comparing them, so that two tarballs, ISO images, PDF documents or software packages can be compared as easily as text files.

It supports a very large list of formats, including APK, Debian and RPM packages, ELF and PE binaries, Java class files, file system images, databases, images, office documents and compressed files. When a format is unknown it falls back to a hexdump comparison, and fuzzy matching handles renamed files. Results are produced as text or HTML.

diffoscope is developed within the Reproducible Builds effort and is free software under the GPL-3.0 licence. It is installed with pip, Docker or most package managers, and can be tried online on try.diffoscope.org.