Bcrypt Hash Generator and Verifier

Features

• Both directions: generate a hash, or verify a password against one
• Adjustable cost factor from 4 to 15, with the rough timing of each shown
• Emits the `$2y$` prefix (PHP-compatible) and accepts `$2a$` / `$2b$` when verifying
• The hash carries its own salt and parameters — nothing else needs saving
• A fresh random salt each time, so the same password hashes differently every run
• A wrong length or format is explained specifically, rather than a bare "verification failed"

How to use

1. Choose the mode: generate or verify
2. Set the cost factor when generating; paste the existing hash when verifying
3. Paste the password into the input box
4. Press Ctrl/⌘ + Enter or click Run, then copy the result

FAQ

What cost factor should I use?

The default of 10 (about 60 ms) suits the vast majority of cases. The trade-off is "how long does my server wait per verification" against "how long does an attacker compute per attempt" — the same number, which is why you cannot simply crank it up. 12 is a common choice; 15 makes each verification approach two seconds and will visibly strain a busy site. For stronger protection, switch to scrypt or Argon2, which can demand more memory at the same wall-clock cost — and memory is exactly the resource attackers struggle to scale.

Why does the same password produce a different hash each time?

Because a random salt is generated every time and embedded in the hash string (the 22 characters after `$2y$10$`). That is what defeats rainbow tables: without a salt, an attacker precomputes hashes of common passwords once and looks up every matching record at once. With a per-record random salt, the same password hashes differently in every row, and that table is useless. So different output on each run is correct behaviour, not a bug.

When verification fails, how do I tell a wrong password from a malformed hash?

This tool distinguishes them. A malformed hash (no `$2y$` prefix, or not 60 characters) reports specifically that it is not a bcrypt hash and shows the expected format. A well-formed hash that simply does not match reports "no match". Also note that copying a hash out of a database often drags in whitespace or a newline — that is trimmed automatically at both ends.

What are bcrypt's limitations?

Two are worth knowing. First, it truncates at 72 bytes, so the tail of a longer password is ignored — "a longer password is safer" does not hold for bcrypt (this tool will not error on it, but you should know). Second, it consumes CPU but not memory, which makes GPU-parallel cracking comparatively cheap — precisely the problem scrypt and Argon2 were designed to fix. For an ordinary website bcrypt remains entirely adequate; where the cryptographic bar is higher, use Argon2id.