Skip to content

MarkLLM

An open toolkit for LLM watermarking

Visit official site

What MarkLLM is

The toolkit is described as an open-source package that gives researchers a unified, extensible way to evaluate LLM watermarking.

What you can do with it

  • Run several watermarking algorithms
  • Visualise how each algorithm behaves
  • Evaluate detectability, robustness and text quality

Pros & cons

✓ What we like

  • A customisable automated evaluation pipeline is listed
  • Test cases and example scripts are provided

! What to watch out for

  • Evaluation covers detectability, robustness and text quality
  • Setup requires a Python environment

FAQ

Which algorithm families?

The material names the KGW family and the EXP family.

How many evaluation tools?

The material states 12 evaluation tools.

Last reviewed: 2026-11-15

More AI coding tools tools

View all →

How we review