Live tool · k-anonymity prefix query

k-Anonymity Password Check

Verify whether a password appears in known public breaches using zero-knowledge k-anonymity. Your password is SHA-1 hashed in your browser; only the first 5 hex characters are sent to the range API. If the database cannot be reached, the tool says so instead of guessing — it will never call a password safe on the strength of a lookup that failed.

Run a k-anonymity breach lookup

Type any password. Your browser computes its SHA-1 hash locally and queries only the 5-character prefix against the Pwned Passwords range API. The password and full hash never leave your device. If the database cannot be reached, you get an unavailable message with the reason and a retry — not a verdict.

Zero-knowledge guarantee. Only a 5-character SHA-1 prefix (~1 million matching hashes per prefix) is requested over HTTPS with padding enabled. If that request fails, you get an unavailable message, not a verdict. A password that appears in a short built-in list of well-known breached passwords is still flagged while offline, because a hit is a hit — it is the “not found” that would prove nothing.

What is a breach?

When a service is compromised, its user database — emails, password hashes, personal details — can be stolen and later published or sold. Billions of such records circulate in "combo lists" used for credential stuffing.

Why reuse is the real danger

Attackers don't need to hack your bank — they take a password leaked from a breached forum and replay it everywhere you have an account. Unique passwords turn a breach into a one-account problem.

The model

How k-anonymity breach lookup works

The privacy-preserving range architecture powered by Have I Been Pwned's Pwned Passwords API.

Hash locally

Your password is hashed on your device using WebCrypto SHA-1. The plaintext password and full 40-character digest never leave your browser.

Send only a 5-char prefix

Only the first 5 hex characters of the hash are queried — a range shared by hundreds of unrelated hashes, with response padding enabled.

Match locally

The range endpoint returns candidate 35-character suffixes; your browser compares them locally in memory and discards the list immediately.

Transparent network behavior: open your browser's Network inspector while running a check — you will see a single GET request containing only 5 hex characters (api.pwnedpasswords.com/range/XXXXX), and never your password. Read the full k-anonymity explainer or our privacy policy.

Why hashing alone doesn't make passwords safe

It is tempting to think "hash it and it's protected". Reality is harsher:

  • Fast hashes are brute-forceable. MD5/SHA-1 are designed for speed; a modern GPU rig tries ~10¹¹ candidates per second against leaked databases.
  • Unsalted hashes leak equality. Identical passwords produce identical hashes, enabling rainbow tables and instant matches across users.
  • Hashing is not encryption. It is one-way only against inputs with enough entropy — human passwords usually don't have it, which is exactly what cracking exploits.
  • The fix is layered: salted, memory-hard hashing (Argon2/bcrypt) on the server side; unique, generated passwords and 2FA on yours.

If a service you use was breached, follow the playbook in What to Do After a Data Breach.

Frequently asked questions

How do I check if my password is leaked?

This page does it without sending the password anywhere: your browser hashes it with SHA-1, asks the breach database for one 5-character prefix, and compares the returned list locally. A count above zero means the password appears in known breaches — change it everywhere it was used.

What should I do if my password is compromised?

Change it on every account that used it, starting with your email — that is the key to password resets everywhere. Turn on 2FA, and replace reused passwords with unique generated ones so the next breach costs one account instead of all of them.