Pepper

What it means

A pepper is a secret value mixed into password hashing and stored somewhere other than the database. It is applied to every password, usually by running the hash through a keyed function such as HMAC, or by encrypting the hash. Without the pepper, a thief holding only the stolen database cannot test password guesses.

It is often confused with a salt. A salt is stored next to each hash, is unique per user, and is not secret. A pepper is the same for everyone, is secret, and lives in a different place: a secrets manager, a hardware module or protected application configuration.

Why it matters

Most database thefts come through a flaw in the application or the data store, such as an injection bug. The attacker leaves with the hashes and starts guessing offline. If the pepper sits in a different system, they are missing a key ingredient, and the hashes cannot be cracked until it is also found.

NIST’s guidance on digital identity recommends this extra keyed step, using a secret key known only to the verifier and stored separately, ideally in hardware. It is a second layer behind slow hashing, not a replacement for it, and it only helps if the key really is kept apart from the data.

How it shows up

In practice, the application reads the pepper at start-up from a secret store, applies it to the password or the hash, and then runs the normal password hashing. The pepper never goes in the database, in source control or in logs. A stored version number next to each hash lets you know which pepper was used.

The hard part is rotation. Changing the pepper invalidates every stored hash unless you keep old versions or re-wrap hashes at the next login. Other mistakes: committing it to the repository as a hardcoded secret, keeping it in the same database table, or treating it as a substitute for a slow algorithm.