Skip to content

ZipPy

Estimate text perplexity through compression ratios

Visit official site

What ZipPy is

The tool is described as a fast research detector that measures text perplexity indirectly through compression ratios. It compares a provided sample with AI generated corpora and returns a statistical comparison.

What you can do with it

  • Install the package or build it from source
  • Pass a text file through command line arguments
  • Import it as a Python package in a script

Application scenarios

  • Moderation teams filter AI generated news
  • Academic institutions screen submitted papers
  • Businesses watch feedback for chatbot replies

Pros & cons

✓ What we like

  • Can be imported as a Python package in other scripts
  • Output is a statistical comparison with known LLM corpora

! What to watch out for

  • Classification relies on similarity to the AI corpus
  • The result is a statistical comparison rather than a verdict

FAQ

How is it installed?

The guide lists pip3 install thinkst-zippy and a source build.

Which compressors are used?

LZMA, Brotli and zlib compression ratios are named.

Last reviewed: 2026-11-15

More AI text detection tools

View all →

How we review