Back to updates
New releaseNov 17, 2026

remove-ai-watermarks source-classify-v7-freeze-2026-11-16

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

Share

Remove AI Watermarks

Remove AI provenance marks from images, video, and audio you generated yourself:

  • known visible labels such as the Google Gemini sparkle watermark and vendor text marks;
  • invisible pixel watermarks through direct local-format disruption or diffusion regeneration;
  • compatible public audio watermarks, with source and saved-output decoder checks;
  • C2PA, EXIF, XMP, IPTC, and related AI metadata.
BeforeAfter
Image with a visible Gemini watermarkImage after visible watermark removal
uv tool install "remove-ai-watermarks[visible]"   # CLI for visible marks and metadata
npx skills add wiltodelta/remove-ai-watermarks    # agent skill for Claude Code, Codex, Cursor, and other agents

Video support covers provenance identification, complete visible-plus-metadata cleaning, directory batches, visible Sora, Veo, Seedance, Doubao, Dola, Hailuo AI, and Kling AI mark removal, and oracle-certified VAE regeneration for video SynthID removal.

raiw.cc runs this library as a hosted service, with the GPU included and nothing to install. Visible mark and metadata removal at Standard output up to 12 MP are free there; original resolution above 12 MP and invisible watermark removal are paid.

PyPI Python Downloads License Tests Sponsor skills.sh Glama MCP server

This is a research project on AI provenance signals, for lawful use on content you own. It is provided as is, without warranty or liability, and each user is responsible for how they use it. It does not target stock agency previews or other watermarks that protect third party paid content. Some jurisdictions, including China, restrict removing AI labels or supplying tools that do. See scope, safety, and legal notes.

Choose what you want to do

GoalCommandGPU
Find provenance signals and watermarksidentifyNo
Classify likely AI provider and model version from pixels (opt-in, not provenance)classify-sourceNo
Remove known visible AI marksvisibleNo
Erase a region you selecteraseNo
Strip AI metadatametadataNo
Identify supported video provenancevideo identifyNo
Inspect audio metadata and compatible public-model evidenceaudio identify (--scheme auto, audioseal, wavmark, perth, silentcipher, silentcipher16, or all)No
Strip AI metadata from audioaudio metadataNo
Process a detected compatible audio watermark and verify the saved file (.wav, .flac, .mp3, .m4a, .ogg, .opus)audio invisibleNo
Strip AI metadata and optionally clean compatible audio watermarksaudio all / audio batchNo
Process one or all audio tracks (--all-streams) while copying encoded videovideo audioNo
Remove visible marks, AI metadata, and opt-in video/audio invisible marksvideo allDepends on --invisible
Strip AI metadata from videovideo metadataNo
Remove a registered visible AI mark from videovideo visibleNo
Process a directory of videosvideo batchDepends on mode
Inspect and disrupt a known public VideoSeal/PixelSeal message (experimental)video messageOptional
Regenerate video pixels (--profile auto, calibrated SynthID, Seedance, or native CUDA Wan 2.1)video invisibleRecommended
Identify C2PA text/HTML wrappers, VSRMark, and Unicode steganographytext identifyNo
Strip C2PA text/HTML wrappers, VSRMark, and Unicode steganographytext metadataNo
Regenerate unlocked prose for best-effort token-watermark removaltext invisibleRecommended
Strip text provenance and optionally regenerate prose spanstext allDepends on --invisible
Regenerate an image to disrupt invisible watermarksinvisibleRequired (CUDA)
Run visible, invisible, and metadata removalallRecommended
Process a directorybatchDepends on mode

Install remove-ai-watermarks[text] for the default local Transformers text rewriter and its packaged language checker. Local text regeneration also supports optional Apple silicon MLX models through remove-ai-watermarks[text-mlx] and --rewriter-backend mlx --rewriter-model /path/to/checkpoint. Both backends share the same acceptance guards; model output still needs factual review. See the text CLI.

Experimental public VideoSeal and PixelSeal known-message video checks and processing use video message. They require an independently known encoder message and do not certify arbitrary unknown marks or provider deployments.

Microsoft Paint and Photos InvisMark declarations are routed automatically to pixel regeneration. The all command removes both the hidden pixel watermark and its linked C2PA manifest; metadata stripping alone removes only the manifest. A specialized Python API can inspect and disrupt the validated local Watermarker.dll payload without diffusion.

Installation modes

Categories