Pukeko

Pukeko goes in the wild, create tailored wordlists and enumerate credentials from a local folder

Tired of using sort input.txt | uniq > output.txt I wanted to create a cross OS script that could read any possible file, take each word once, and list them all in a word-list.

Requirements

Python packages (required)

Package	Purpose	Install
python-magic	Detect plain text files by content	pip install python-magic
pyxtxt	Extract text from all document and image formats	pip install pyxtxt
openai-whisper	Transcribe audio and video locally	pip install openai-whisper
colorama	Colour terminal output on all platforms	pip install colorama

Install all at once:

pip install python-magic openai-whisper colorama pyxtxt

For broader format coverage install pyxtxt with optional extras:

pip install "pyxtxt[pdf,docx,presentation,spreadsheet,html,ocr]"

System tools (required for certain file types)

Tool	Purpose	Without it
Tesseract OCR	Extract text from images (.jpg, .png, .gif, .tif)	Images will be skipped
ffmpeg	Decode audio and video for Whisper	Audio/video will not work

Pukeko will warn you at startup if any system tool is missing.

How to Install

1. pip install python-magic openai-whisper colorama pyxtxt

2. Install Tesseract OCR for image support

3. Install ffmpeg for audio/video support

Troubleshooting:

• python-magic: as if the World wasn't complicated enough, there are 2 'Magic' libraries. You can find the right one here on GitHub or here on pypi.python.org
• openai-whisper: runs fully locally — no API key, no internet, no token limits. See the GitHub repo for details. A GPU is optional but speeds up transcription significantly.
• pyxtxt: supports many formats out of the box; install with extras (pyxtxt[ocr], pyxtxt[pdf], etc.) for broader coverage. See the pyxtxt PyPI page for the full list.

If the situation gets tragic open an issue and I will help you troubleshooting

How to use it

Pukeko can currently parse:

Documents & images (via pyxtxt): '.csv', '.doc', '.docx', '.eml', '.epub', '.gif', '.htm', '.html', '.jpeg', '.jpg', '.json', '.log', '.msg', '.odt', '.pdf', '.png', '.pptx', '.ps', '.psv', '.rtf', '.tff', '.tif', '.tiff', '.tsv', '.txt', '.xls', '.xlsx'

Audio & video (via openai-whisper, transcribed locally): '.mp3', '.wav', '.ogg', '.flac', '.m4a', '.aac', '.wma', '.mp4', '.avi', '.mkv', '.mov', '.wmv', '.flv', '.webm', '.m4v'

Plain text files (via python-magic): any file identified as plain text by the system, e.g. .py, .js, .xml, shell scripts, config files, and other text-based formats.

The -model flag lets you choose the Whisper model size for audio/video transcription:

Model	Speed	Accuracy
tiny	fastest	lowest
base	fast	low
small	balanced	good (default)
medium	slow	very good
large	slowest	best

Example: python Pukeko.py -input /path/to/files -output wordlist.txt -model medium

Have a look at my YouTube presentation:

Future developent

On spare time my TODO list is:

• add option -URL to create wordlists from a target web page like CeWL
• add option -site to create wordlists from a target website
• add option Leet (or 1337), also known as eleet or leetspeak (so many passwords are week because of leetspeak )
• add multilanguage (pip install alphabet-detector)
• add highlight HotWords in string
• add e-mail to HotWords