See — and remove — C2PA Content Credentials.
AI generators and Adobe tools now embed signed provenance manifests in your images. CleanImages shows you what's inside the manifest, then removes it — entirely in your browser.
- 100% in your browser
- No uploads, ever
- Reads the manifest before removing it
- Free
CleanImages tool
100% IN YOUR BROWSER · NO UPLOADS, EVER · NO ACCOUNT REQUIRED. Try it offline. Disconnect your network and reload — it still works.
Your image may be carrying a signed receipt.
If a file was generated by ChatGPT, Gemini, or Firefly — or touched by Photoshop's AI features — it likely carries a C2PA manifest: a cryptographically signed block recording what tool made it, when, and what operations were applied. Platforms like Instagram and TikTok read it on upload and can label your post based on it.
Sometimes that's exactly what you want. When it isn't, you should be able to see it and remove it. Both, here, free.
Look before you strip.
The report shows who signed the manifest (Adobe, OpenAI, …), what it asserts (AI-generated, edit history), and any data embedded inside the signed block — including thumbnails and EXIF that survive naive EXIF-only cleaners.
What removal does — and what it can't.
Removing a C2PA manifest removes the readable, machine-checkable metadata declaration from your file. It does not remove pixel-level watermarks like Google's SynthID, which survive metadata removal entirely — and it doesn't make AI images undetectable to pixel-based classifiers. If a tool tells you otherwise, it's selling you the wrong half of the truth.
Also worth knowing: a manifest is an authenticity signature. If you need your audience to verify how an image was made — journalism, client work with disclosure expectations — keep it. We flag before removing one, so the choice is deliberate.
- Is it legal to remove C2PA metadata?
- The EU AI Act's marking rules bind AI providers, not end users cleaning their own files — but deceptive uses (deepfakes presented as real) carry their own obligations. This isn't legal advice.
- Does this remove the “AI info” label on Instagram?
- The label is triggered by metadata at upload, so cleaning before you post removes that trigger. It doesn't remove labels from posts already up.
- Which tools embed C2PA?
- ChatGPT/DALL·E, Google Gemini/Imagen, Adobe Firefly, and Photoshop's AI features, among others. Midjourney currently doesn't.
- Will removing it change my image?
- No. The manifest lives in the metadata section of the file; removing it leaves the compressed image data untouched.
Check what your image is carrying — free, in your browser, nothing uploaded.
Clean an image