# wiltodelta/remove-ai-watermarks

Remove visible and invisible AI watermarks and provenance metadata from images and video. Python library and CLI for SynthID, C2PA, EXIF, IPTC, XMP, and common generative-AI marks.

Repository: https://github.com/wiltodelta/remove-ai-watermarks
Canonical: https://ross.abutalabs.com/products/remove-ai-watermarks
Homepage: https://raiw.cc
Language: Python
License: Apache-2.0
License Family: permissive
Topics: cli, computer-vision, diffusion-models, image-processing, metadata, python, watermark-removal, c2pa, gemini, synthid, ai-watermark, content-credentials, exif, flux, generative-ai, nano-banana, stable-diffusion, comfyui, watermark-remover, video-processing
Last push: 2026-08-26T18:22:38+00:00

## Health v2 (maintenance only)
Score: 77/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 87, longevity 11
- inputs: {"age_days": 161, "days_push": 7, "days_rel": 6, "gap_med": 0.0, "n_releases_24m": 79}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5278, forks 496 (observed 2026-08-28T04:09:14.062533+00:00)

## What it is
A Python library and CLI for removing AI watermarks and provenance metadata from images and video the user generated themselves. It handles visible vendor marks (Gemini sparkle, Sora, Veo, Kling, etc.), invisible pixel watermarks like SynthID via diffusion/VAE regeneration, and metadata such as C2PA, EXIF, XMP, and IPTC.

## Use cases
- remove the Gemini sparkle watermark from my AI-generated image
- strip C2PA content credentials and 'Made with AI' metadata from photos
- remove SynthID invisible watermark from images I generated
- clean visible AI marks from Sora or Veo videos
- batch process a directory of AI images to remove watermarks and metadata
- identify what AI provenance signals exist in an image or video
- erase a selected region of an image to remove a vendor label

## When to choose
- you need to clean provenance marks from AI content you own, across visible, metadata, and hidden pixel layers
- you want a scriptable CLI or Python API with batch processing for images and video
- you need SynthID removal via GPU-based regeneration with face restoration

## When to avoid
- you want to remove watermarks from third-party or stock content you do not own - the project explicitly targets only your own generated content
- you have no GPU and need invisible watermark removal, which requires CUDA
- you need a fully hosted no-install solution - the companion site raiw.cc offers that instead

## Facets
- artifact type: library
- maturity: active
- function: image-processing, video-processing, computer-vision, cli, privacy
- domain: image-processing, privacy, artificial-intelligence, developer-tools
- platform: python, cli, windows
- tags: watermark-removal, synthid, c2pa, content-credentials, exif, xmp, iptc, metadata-stripping, generative-ai, diffusion-models, stable-diffusion, gemini, provenance, invisible-watermark, video-watermark, metadata, video, command-line, linux, macos, gpu

## Member repositories
- wiltodelta/remove-ai-watermarks (main) score 77

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:14.062533+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:59:18.103164+00:00, confidence not recorded.
  - readme: https://github.com/wiltodelta/remove-ai-watermarks (fetched 2026-08-28T04:09:14.062533+00:00, sha 70e7012faa18)
  - homepage: https://raiw.cc (fetched 2026-08-29T08:54:11.736962+00:00, sha 117e32d817a2)
  - registry_pypi: https://pypi.org/pypi/remove-ai-watermarks/json (fetched 2026-08-29T08:54:11.746562+00:00, sha a9b894c72202)
- Data as of 2026-08-30T08:39:29.467469+00:00.