Password guides

Compare a predictable password with an unrelated-word passphrase

·

A familiar word with capital letters, digits and punctuation can still follow a common password pattern. A pattern-based strength checker estimates guessability by looking for dictionary matches, repeats and other familiar structures. Compare the public examples to see how the score responds.

Password Strength Checker

See how familiar words and sequences affect the meter

Choose Load predictable example and then Check strength. Password123! is 12 code points long. The model gives it 1 out of 4 and reports a dictionary or common password match.

Try Load passphrase example and run the check again. The phrase tundra violet locket osprey gives 4 out of 4. The model can still recognize dictionary words. The way the words combine raises its estimate. These are public examples for comparison. Use independently chosen credentials for accounts.

See how familiar words and sequences affect the meter
Public test sampleCode pointsModel score
Password123!121 / 4
tundra violet locket osprey274 / 4

zxcvbn-ts guide, score and feedback fields

Read the score as a model estimate

Try 1234512345. The estimator recognizes the repetition and shows 0 out of 4. Read the pattern feedback to see which part of the choice is easy to predict. For instance, adding another copy of the same sequence preserves the underlying pattern.

The number of guesses in the model describes the supplied string. A real service can rate-limit attempts or use a slow password hash. Password reuse, leaked credentials and an attacker’s knowledge of personal facts can affect the outcome. The checker evaluates the input against bundled dictionaries and patterns without consulting any account or breach service.

zxcvbn-ts guide, score and feedback fieldsNIST SP 800-63B-4, strength of passwords

Check which length a password policy measures

Enter é🙂 and run the check. The display counts two Unicode code points and six UTF-8 bytes. Different storage schemes count different units. A site can also impose its own character rules.

The model uses common and English dictionaries. A high score offers weaker evidence for words or personal patterns outside that data. Where a legacy system uses bcrypt, check the UTF-8 byte limit separately.

zxcvbn-ts guide, score and feedback fieldsNIST SP 800-63B-4, strength of passwords

Change the checklist and keep the same model score

Keep Check as I type enabled and load the predictable example. At a 15-code-point minimum, Password123! meets four of six checklist items. It fails the length check and contains adjacent s characters. The pattern model shows 1 out of 4.

Choose 12 code points in Minimum length to check. The same sample now meets five of six checklist items, while its model score stays 1 out of 4. Match the checklist to the service’s published rules. NIST’s single-factor guidance sets a 15-character minimum for its scope and advises against requiring mixtures of character types.

Open Password Strength CheckerNIST SP 800-63B-4, strength of passwordszxcvbn-ts guide, score and feedback fields

Generate and check a fresh candidate on your device

Open Generate a random password. Set Generated password length to 20 and keep the character sets your service accepts. Choose Generate password. The new candidate appears in the masked field, and the local model evaluates it when live checking is enabled.

A modeled guess duration at a billion guesses per second is arithmetic applied to that model. Actual guessing time depends on the service’s rate limits and storage method. Use Copy generated password only when you want to move the candidate to your password manager, then clear the field.

Open Password Strength CheckerW3C Web Cryptography API random valuesNIST SP 800-63B-4, strength of passwords

Common questions

Does a 4 out of 4 score guarantee account protection?

A high score estimates lower guessability against the bundled model. Unique credentials, service rate limits, password storage and multi-factor authentication also matter.

Can this checker estimate how long a real attacker would take?

The tool reports a modeled number of guesses. A real duration also depends on access, rate limits, hardware and the password hashing algorithm.

Sources

Put it into practice.

Related guides