gwpgen.py

gwpgen.py is an entropy-driven, cryptographically sound password generator written in Python.

It generates strong random passwords using the operating system's cryptographically secure random number generator and calculates password length from a desired entropy target.

Features

• cryptographically secure randomness (os.urandom)
• entropy-driven password length calculation
• no modulo bias (rejection sampling)
• Fisher–Yates shuffle for unbiased character ordering
• configurable character classes
• optional grouping for human-friendly passwords
• WiFi QR string output
• entropy presets for common security levels
• optional enforcement of character classes
• configurable safety limits

Installation

Clone the repository:

git clone https://github.com/georgwallisch/gwpgen.git
cd gwpgen

Run directly:

$ ./gwpgen.py

Or install it somewhere in your $PATH.

Basic usage

Generate a strong password with 128 bits of entropy:

$ ./gwpgen.py --strong -a -m

Example output: kF7!dP3$wQ9r#N8TzB6M

Generate a password of fixed length:

$ ./gwpgen.py -L 20 -a -m

Generate multiple passwords:

$ ./gwpgen.py --strong -a -m -N 5

Group characters for better readability:

$ ./gwpgen.py --strong -a -m -g 4

Example: vT8d-9Xk3-ZP4q-Lm2B

Pad the last group to full group size:

$ ./gwpgen.py --strong -a -m -g 4 -p

Character classes

You must select at least one character class.

Option	Description
-a, --alphanum	Uppercase, lowercase and digits
-u, --upper	Uppercase letters
-l, --lower	Lowercase letters
-d, --digits	Digits
-m, --symbols	Symbols (optional custom set)
-x, --hex-lower	Hexadecimal lowercase
-X, --hex-upper	Hexadecimal uppercase

$ ./gwpgen.py --strong -u -l -d -m

Entropy presets

Preset	Entropy
--weak	50 bits
--good	70 bits
--strong	100 bits
--stronger	128 bits
--tough	192 bits
--paranoid	256 bits
--insane	512 bits

Example:

$ ./gwpgen.py --paranoid -a -m

Entropy vs length

Instead of specifying a password length, you can specify a target entropy:

$ ./gwpgen.py -b 128 -a -m

The program will automatically calculate the required password length based on the size of the selected alphabet.

WiFi QR code strings

gwpgen can output passwords in WiFi QR format compatible with Android and many QR code generators.

Example:

$ ./gwpgen.py --strong -a -m -W MyNetwork

Output example: WIFI:T:WPA;S:MyNetwork;P:abcDEF123!;

Hidden network:

$ ./gwpgen.py --strong -a -m --wifi-hidden MyNetwork

Output options

Write passwords to a file:

$ ./gwpgen.py --strong -a -m -o passwords.txt

Write both to file and stdout:

$ ./gwpgen.py --strong -a -m -o passwords.txt --tee

Safety limits

To avoid accidental misuse, gwpgen enforces limits by default:

• entropy 1000 bits
• length 250 characters
• number of passwords 500

Disable limits:

$ ./gwpgen.py --no-limits

Security properties

gwpgen was designed to avoid common pitfalls found in many password generators.

Security features include:

• operating system CSPRNG (os.urandom)
• rejection sampling to avoid modulo bias
• Fisher–Yates shuffle
• entropy-based password generation
• optional character class enforcement
• configurable alphabet

Example

Generate five strong passwords grouped for readability:

$ ./gwpgen.py --strong -a -m -g 4 -N 5

Example output:

T4p!-9Xm2-rQ6d-Lz7A 3Kq%-Lm9P-x2fD-aT8Z hV3#-F8kW-rP2q-M9bC 8Qw@-sF3T-vX9d-H4mL N7x!-q2pD-KM9f-Wz3R

Short options can be grouped together:

$ ./gwpgen.py --strong -vapg

Equivalent to:

$ ./gwpgen.py --strong -v -a -p -g

Generate a numeric password for embedded devices such as pay terminals using a dot as separator:

$ ./gwpgen.py -L 24 -dpg 4 -s "." -o mywifi.txt --tee

Example output: 3109.4236.2589.1904.1843.7841

Generate a strong password for some login:

$ ./gwpgen.py --strong -apg 4

Example output: r3s6-y9O6-xYRd-tylt-Z5lW

or stronger:

$ ./gwpgen.py --stronger -apg 4

Shorthand method: Group size default 4 and -G combines -g and -p:

$ ./gwpgen.py --stronger -aG

Example output: OLS5-9YKM-291c-vfv7-su5h-38Sx

or with symbols and not grouped:

$ ./gwpgen.py --stronger -am

Example output: 7a&b+GoJy(SNV;RdlM53M

License

Copyright (c) 2026 Georg Wallisch

MIT License – see LICENSE