Topics / #c2pa

C2pa projects on GitHub

Content provenance and watermark standards, including tools that strip AI watermarks.

As of , the most-starred C2pa projects on GitHub are watermarks-remover (24.4K stars), remove-ai-watermarks (6K stars) and Turnitout-Humanizer (2K stars). GitHub.Locker tracks 5 C2pa repositories with 32.9K stars combined. The fastest-growing on October 18 was Turnitout-Humanizer, with +1.7K stars in a day.

5Projects
32.9KStars in total
+1.8KStars in a day

Hot right now

Most stars gained in a day · Oct 18

The all-time top

5 projects · by stars

  1. #1 · guillaumemeyer/watermarks-remover

    watermarks-remover

    It strips AI provenance marks from content you own — invisible Unicode tricks in text, statistical text watermarks, and C2PA/EXIF/XMP metadata in files — via an agent skill plus a local Python service, privacy-first.

    24.4K
  2. #2 · wiltodelta/remove-ai-watermarks

    remove-ai-watermarks

    A Python library and CLI that removes visible and invisible AI watermarks and provenance metadata (SynthID, C2PA, EXIF, IPTC, XMP, and vendor marks like Gemini's sparkle) from images and videos you generated yourself.

    6K
  3. #3 · AhmadHassan-BTed/Turnitout-Humanizer

    Turnitout-Humanizer

    Turnitout is a rule-based engine that strips AI-detector watermarks and plagiarism-similarity signals from documents and code so they pass Turnitin, iThenticate, GPTZero, CopyLeaks, and ZeroGPT, without using any generative AI itself.

    2K
  1. 4

    noai-watermark

    209mertizci/noai-watermark · Python

    noai-watermark is an open-source CLI and Python toolkit that uses diffusion-based image regeneration to disrupt invisible AI watermarks (like Google SynthID, StableSignature, TreeRing) embedded in generated images, and it also inspects/strips AI metadata (EXIF, PNG text chunks, C2PA manifests).

  2. 5

    anthropies

    191CharlesHoskinson/anthropies · TypeScript

    Anthropies is a CLI tool and local HTTP service that strips vendor marks (invisible Unicode, trailers, banners, C2PA and file metadata) from AI-generated outputs you already own. It explicitly does not remove Anthropic's core keyed text watermark — only the metadata and formatting layers around it.

Often tagged together