watermarks-remover

An agent skill and local Python service that strips AI provenance marks, such as hidden characters and file metadata, from text and files you own. It is meant for privacy and cleanup.

Share on XLicense: MIT

Overview

watermarks-remover is an agent skill plus a local Python service that strips AI provenance marks from text and files you own, for privacy and cleanup. It removes invisible Unicode characters and unusual spaces from text, and metadata such as C2PA, EXIF and XMP from many file types. The skill is a thin client that talks to the service over HTTP, so the agent host needs no Python.

Key features

  • Deterministic scripts remove invisible Unicode, bidi and tag characters
  • Cleans C2PA, EXIF, XMP and document metadata in many formats
  • Supported files include images, PDF, Office documents, audio and video
  • Agent rewrite option for statistical text watermarks
  • Skill installs into Claude Code, with a plugin marketplace route

Best for

People who want to clean provenance data from their own content before sharing it. It is meant for content you own, and the README frames it around privacy and hygiene.

Upstream
guillaumemeyer/watermarks-remover
Fork on GitHub
Guo-astro/watermarks-remover
Upstream stars
23k
Category
Documents, media and content
Language
Python
License
MIT
Forked
2026-08-16
Sync status
In syncLast synced 2026-09-29