Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
A reproducible build produces bit-for-bit identical output every time it is run from the same source, so an independent party can rebuild a release and prove nothing was inserted along the way.

What Is a Reproducible Build?
A build is reproducible when running it again from the same source, with the same declared build environment and instructions, yields byte-for-byte identical artefacts. Not functionally equivalent. Identical, down to the hash.
The point is verification by someone other than you. If a release is reproducible, an independent party can rebuild it from published source and compare hashes. A match means the binary corresponds to the source. A mismatch means something happened between the two that nobody declared.
Code review, dependency scanning, and signing all operate on the assumption that the artefact you ship is the artefact your source describes. A compromised build system breaks that assumption silently, and it is a well-attested attack: the SolarWinds Orion compromise disclosed in December 2020 inserted the SUNBURST backdoor during compilation, leaving the source repository clean and every signature perfectly valid.
Reproducibility does not prevent that. It makes it detectable, because independent rebuilders comparing their output against the published artefact would have found a divergence they could not explain. This is why the concept keeps appearing alongside SLSA and signed provenance rather than competing with them: signatures tell you who built it, reproducibility tells you whether the build was honest.
Nearly every build is non-reproducible by default, and the causes are mundane:
/home/alice/src and /builds/42 produce different output.The fixes are well understood and mostly cheap. Honour SOURCE_DATE_EPOCH, the environment variable the Reproducible Builds project defined for exactly this purpose, so tools embed a fixed timestamp drawn from the last source change. Sort file lists explicitly. Use deterministic archive settings, or run strip-nondeterminism over the result. Normalise paths — Go’s -trimpath flag, or the equivalent compiler options for C and C++ and Rust. Pin everything: dependency versions by lock file, base images by digest, toolchains by exact version.
Two tools make the work possible. diffoscope compares two builds and explains the difference in human terms, recursing into archives, binaries, and package formats rather than reporting that two files differ. reprotest rebuilds in deliberately varied environments — different paths, locales, timezones, user identifiers — to find assumptions you did not know you had.
The Reproducible Builds project started with Debian in 2013, and the overwhelming majority of Debian source packages now build reproducibly, with public rebuild infrastructure that checks continuously. Tails, Tor Browser, Arch Linux, NixOS, Guix, F-Droid, and Bitcoin Core all rely on independent rebuilds as a core security property, in several cases requiring multiple signers to agree on the hash before a release ships. Go has produced reproducible builds from its toolchain since version 1.21, and reproducibility is an explicit goal of several language ecosystems.
So it is achievable. It is also unevenly achievable: a Go service is usually reproducible after an afternoon of work, while a large C++ application with a bespoke build system can take months.
Be honest about where the value is. For software distributed widely, especially security software, firmware, or anything whose users cannot inspect the build, reproducibility is one of the few ways to offer verifiable assurance. For an internal web service rebuilt from a pinned pipeline and deployed by digest, the marginal gain over good provenance and a locked-down builder is modest.
A reasonable middle path for most organisations: pin dependencies and base images by digest, set SOURCE_DATE_EPOCH, produce signed provenance describing exactly how the build ran, and measure how close to reproducible you already are with diffoscope. Many builds are two or three small changes away, and knowing the distance is worth the afternoon it takes to find out.
At Privia Security, we assess build and release pipelines against this class of supply chain attack as part of our cybersecurity consultancy work.
Because build tools record things that change between runs: timestamps, the directory the build ran in, the hostname, the locale, the order the filesystem returned files in, and any randomness used during compilation. None of it affects behaviour, so it goes unnoticed until someone compares hashes. Fixing it is largely a matter of pinning those inputs and normalising the ones that remain.
No, it makes one detectable. If a build server is compromised and inserts code, the published artefact will not match what an independent rebuild produces, and the discrepancy is visible to anyone who checks. That only helps if someone actually rebuilds and compares, which is why reproducibility is usually paired with independent rebuilder infrastructure or multiple signers who must agree on the hash.
A standard environment variable, defined by the Reproducible Builds project, that tells build tools which timestamp to embed instead of the current time. It is normally set to the date of the most recent source change. Many compilers, archivers, and documentation generators honour it directly, which makes it one of the cheapest and most effective single changes towards a deterministic build.
Use diffoscope. It compares two files or directories recursively, unpacking archives, disassembling binaries, and decoding package formats so it can tell you that a particular object file differs by an embedded build path rather than simply reporting that the bytes are not equal. Pair it with reprotest, which rebuilds under deliberately varied conditions to reveal hidden environmental assumptions.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.