
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.
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.
| Before | After |
|---|---|
![]() | ![]() |
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.
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
| Goal | Command | GPU |
|---|---|---|
| Find provenance signals and watermarks | identify | No |
| Classify likely AI provider and model version from pixels (opt-in, not provenance) | classify-source | No |
| Remove known visible AI marks | visible | No |
| Erase a region you select | erase | No |
| Strip AI metadata | metadata | No |
| Identify supported video provenance | video identify | No |
| Inspect audio metadata and compatible public-model evidence | audio identify (--scheme auto, audioseal, wavmark, perth, silentcipher, silentcipher16, or all) | No |
| Strip AI metadata from audio | audio metadata | No |
Process a detected compatible audio watermark and verify the saved file (.wav, .flac, .mp3, .m4a, .ogg, .opus) | audio invisible | No |
| Strip AI metadata and optionally clean compatible audio watermarks | audio all / audio batch | No |
Process one or all audio tracks (--all-streams) while copying encoded video | video audio | No |
| Remove visible marks, AI metadata, and opt-in video/audio invisible marks | video all | Depends on --invisible |
| Strip AI metadata from video | video metadata | No |
| Remove a registered visible AI mark from video | video visible | No |
| Process a directory of videos | video batch | Depends on mode |
| Inspect and disrupt a known public VideoSeal/PixelSeal message (experimental) | video message | Optional |
Regenerate video pixels (--profile auto, calibrated SynthID, Seedance, or native CUDA Wan 2.1) | video invisible | Recommended |
Identify C2PA text/HTML wrappers, VSRMark, and Unicode steganography | text identify | No |
Strip C2PA text/HTML wrappers, VSRMark, and Unicode steganography | text metadata | No |
| Regenerate unlocked prose for best-effort token-watermark removal | text invisible | Recommended |
| Strip text provenance and optionally regenerate prose spans | text all | Depends on --invisible |
| Regenerate an image to disrupt invisible watermarks | invisible | Required (CUDA) |
| Run visible, invisible, and metadata removal | all | Recommended |
| Process a directory | batch | Depends 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.

