A hash function turns data of any length into a fixed-length fingerprint. The same input always gives the same output, yet changing a single character produces a completely different result, and the original cannot be recovered from the hash. That makes hashes useful for verifying downloads, checking whether two files are identical, de-duplicating database records and building cache or version identifiers.
The tool supports five algorithms: MD5, SHA-1, SHA-256, SHA-384 and SHA-512. You can type text or drop in a file. MD5 is computed with spark-md5 and the SHA family with the browser’s built-in Web Crypto API. Output can be shown in upper- or lower-case hexadecimal, so it is easy to compare character by character with a published checksum. Everything is calculated locally and nothing is uploaded.
Pick the algorithm by purpose. MD5 (128-bit) and SHA-1 (160-bit) can both be made to collide on purpose, so they should no longer be used for signatures or security checks, although they are still fine for spotting accidental corruption in transit. For anything security-related use SHA-256 or stronger; software download pages, Linux images and blockchains widely rely on SHA-256.
A common misconception: hashing is not encryption, and a plain hash is not suitable for storing passwords. SHA-256 is so fast that attackers can try billions of guesses per second. Websites should store passwords with salted, deliberately slow algorithms such as bcrypt, scrypt or Argon2. When verifying a download, only an exact match with the officially published value shows the file has not been altered.