Skip to main content

A Python wrapper for Troy Hunt's Pwned Passwords API.

Project description

CI Status Version Python Versions

pwnedpasswords is a small Python wrapper and command line utility that lets you check if a passphrase has been pwned using the Pwned Passwords v2 API. All provided password data is k-anonymized before sending to the API, so plaintext passwords never leave your computer.


Pwned Passwords are more than half a billion passwords which have previously been exposed in data breaches. The service is detailed in the launch blog post then further expanded on with the release of version 2. The entire data set is both downloadable and searchable online via the Pwned Passwords page.


pwnedpasswords is available for download through PyPi. You can install it right away using pip.

pip install pwnedpasswords


import pwnedpasswords

pwnedpasswords.check("testing 123")
# Returns 1

Security Note

No plaintext passwords ever leave your machine using pwnedpasswords.

How does that work? Well, the Pwned Passwords v2 API has a pretty cool k-anonymity implementation.


Formally, a data set can be said to hold the property of k-anonymity, if for every record in a released table, there are k − 1 other records identical to it.

This allows us to only provide the first 5 characters of the SHA-1 hash of the password in question. The API then responds with a list of SHA-1 hash suffixes with that prefix. On average, that list contains 478 results.

People smarter than I am have used math to prove that 5-character prefixes are sufficient to maintain k-anonymity for this database.

In short: your plaintext passwords are protected if you use this library. You won't leak enough data to identity which passwords you're searching for.


pwnedpasswords automatically checks if your provided input looks like a SHA-1 hash. If it does, it won't do any further processing. If it looks like plain text, it'll automatically hash it before sending it to the Pwned Passwords API.

If you'd like to provide an already-hashed password as input to pwnedpasswords, you don't need to do anything--pwnedpasswords will detect that it looks like a SHA-1 hash and won't hash it again before sending it to the range endpoint.


Likewise, if a password looks like a SHA-1 hash (i.e., matches the regex [0-9a-fA-F]{40}) but is actually a user-provided password, set plain_text to True, so that the library knows to hash it before sending it to the API.

pwnedpasswords.check("1231231231231231231231231231231231231231", plain_text=True)



This is the preferred method. By default, the check method uses the endpoint, which is k-anonymous.

# 34729

If you'd like to force pwnedpasswords to use the search endpoint instead (, set the anonymous parameter to False.

pwnedpasswords.check("password", anonymous=False)
# 3303003

You might want to do this if you'd prefer faster response times, and aren't that worried about leaking passwords you're searching for over the network.

If you'd like direct access to the search and range endpoints, you can also call them directly.


# outputs a dictionary mapping SHA-1 hash suffixes to frequency counts

Command Line Utility

pwnedpasswords comes bundled with a handy command line utility.

$ pwnedpasswords 123456password

Output is simply the number of entries found in the Pwned Passwords database.

If you'd like to prevent input from appearing in your history, specify the --stdin argument to provide input via stdin (h/t to @tveastman for requesting this).

$ pwnedpasswords --stdin

For help, just provide -h as a command-line argument.

$ pwnedpasswords -h
usage: pwnedpasswords [-h] [--verbose] [--plain-text] (--stdin | password)

Checks Pwned Passwords API to see if provided plaintext data was found in a
data breach.

positional arguments:
  password      The password or hashed password to search for.

optional arguments:
  -h, --help    show this help message and exit
  --verbose     Display verbose output.
  --plain-text  Specify that the provided input is plain text, even if it
                looks like a SHA-1 hash.
  --stdin       Read provided input from stdin.


The CLI returns an exit code equal to the base-10 log of the result count, plus 1. If there are no matches in the API, the exit status will be 0. While returning the base-10 log might seem odd, note that most systems require exit status codes to be in the range 0-127, and I wanted the status code to provide some indication for severity. log(N) seemed to be a good tradeoff. The exit status is log(N)+1 since there are plenty of matches in the database with 1 match.

If you'd like to take a look under the hood to make sure things are working as they should, set the --verbose flag.

$ pwnedpasswords 123456password --verbose
INFO:pwnedpasswords.pwnedpasswords:Entry found


Special thanks to Troy Hunt for collecting this data and providing this service.


Dan Loewenherz / (@dlo)

See also

django-pwnedpasswords-validator, a validator that checks user passwords against the Pwned Passwords API using this library.


Apache License, Version 2.0. See LICENSE for details.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release. See tutorial on generating distribution archives.

Built Distribution

pwnedpasswords-2.0.0-py2.py3-none-any.whl (13.6 kB view hashes)

Uploaded py2 py3

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page