mertizci

mertizci / noai-watermark

Public

Remove invisible AI watermarks (SynthID, StableSignature, TreeRing) and strip AI metadata from images. Open-source CLI & Python toolkit.

github.commertizcinoai-watermark ai-watermark c2pa cli computer-vision ctrlregen
33
9
100% credibility
Found Feb 22, 2026 at 22 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A tool for removing invisible watermarks and AI metadata from images generated by tools like Google Gemini, DALL-E, and Stable Diffusion while preserving visual fidelity.

How It Works

1
🖼️ Spot the hidden watermark

You create a beautiful image with an AI tool like Gemini or DALL-E, but it carries invisible marks proving it's AI-made.

2
📦 Grab the cleaning tool

You find and set up this easy image cleaner on your computer to wipe away those hidden signs.

3
📁 Choose your picture

Pick the AI image you want to clean and tell the tool where to save the new version.

4
✨ Magic clean happens

The tool gently refreshes your image to break the hidden watermarks while keeping every detail looking just right.

5
🔧 Tweak for perfection

Adjust the cleaning power or pick a quality mode if you want faster results or the absolute best look.

6
🔍 Check it's gone

Test the new image with detection tools to confirm no traces remain.

✅ Pure image ready

Your picture is now free of AI fingerprints, looking natural and ready to share anywhere.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 22 to 33 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is noai-watermark?

noai-watermark is a Python CLI tool and library that strips invisible AI watermarks like SynthID, StableSignature, and TreeRing from images generated by tools such as Google Gemini, DALL-E, or Midjourney, while also removing AI metadata like prompts and seeds. Run `pip install noai-watermark` then `noai-watermark source.png -o cleaned.png` for quick regeneration that keeps visuals intact but breaks detection. It handles PNG/JPEG via diffusion pipelines on CPU, GPU, or Apple Silicon, with Python APIs for batch processing or metadata cloning.

Why is it gaining traction?

Unlike basic metadata strippers, it targets embedded pixel signals that survive edits or screenshots, using controllable regeneration for near-identical outputs—ideal for noai watermark removal without quality loss. Devs like the dual pipelines (fast default vs high-fidelity CtrlRegen), auto-device selection, and Hugging Face integration, plus online demo for testing. It exposes how fragile these "robust" watermarks are, sparking interest beyond just removal, like pairing with tools to remove invisible characters from AI text.

Who should use this?

Security researchers stress-testing AI provenance systems, or devs building image pipelines needing clean assets free of trackers. Content creators dodging platform AI filters, or teams auditing uploads for hidden markers. Skip if you're just stripping EXIF—grab it for noai watermark scenarios or when remove invisible characters & ai watermarks hits your workflow.

Verdict

Solid early prototype with excellent docs, CLI, and tests—install and verify SynthID removal in minutes despite 14 stars and 1.0% credibility score. Maturity lags (low adoption, CPU-heavy defaults), but it's production-ready for targeted use; watch for GPU optimizations.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.