Privacy · hashHash of a file or text

A short fixed-length string computed from the contents of a file. SHA-256 always produces 64 characters, whether the file is one kilobyte or four gigabytes. Identical files produce identical hashes, while changing a single byte changes the whole result in a way nobody can predict in advance. That is what makes a hash usable as a fingerprint of the contents.

Drop the downloaded file into the field above and paste the reference hash published on the page you downloaded from. A match means the file arrived intact and was not swapped. A mismatch usually means an interrupted download, occasionally a mirror worth avoiding. The algorithm is picked from the length of the reference, and case, stray spaces and a filename pasted alongside do not matter.

No, the file stays on your device. The browser reads it and computes the digest with its built-in Web Crypto module, so neither the file nor the text nor the result is transmitted anywhere. Open the page, disconnect from the network and it keeps working. That is what makes it acceptable to check confidential documents here rather than on a service that wants an upload first.

SHA-256 when the choice is yours: it is enabled by default, it is sound, and it is fast enough for a multi-gigabyte disk image. When you are comparing against someone else's published value, use whatever algorithm that value was produced with, otherwise the strings can never match. You rarely have to pick manually — the length of the pasted reference identifies the algorithm.

Two different files with the same MD5 can be constructed on purpose; for SHA-256 no such method is known. Against accidental corruption during a download both work equally well and differ only in speed. Against deliberate substitution a matching MD5 proves nothing. That is why projects publish SHA-256 today, and MD5 or SHA-1 mostly survive next to older releases.

No. The hash is far shorter than the data and stores none of it: sixty-four characters cannot be unfolded back into a gigabyte image no matter how much computing power is thrown at the problem. A hash function has no inverse operation at all. It answers exactly one question — whether the contents match those the reference was made from.