AI Metadata & C2PA Provenance Checker
Inspect files for C2PA Content Credentials manifests, AI generation parameters, and cryptographic asset provenance.
Inspect AI Signatures & C2PA Manifests
Detects Content Credentials, DALL-E, Midjourney, Adobe Firefly, and Stable Diffusion tags.
Select or drop a file to inspect metadata
Inspect EXIF, GPS, camera serials, PDF author details, Office revisions, audio ID3 tags, video container tags, and C2PA provenance.
Crucial Technical Clarification: Metadata vs Pixel Watermarks
1. Metadata & C2PA Manifests (Removable)
These are digital headers and cryptographic packets stored inside file container boxes (JUMBF, XMP, PNG chunks). Removing metadata strips these tags so automated scrapers cannot read generator prompts or editing history.
2. Pixel-Level Watermarking (NOT Removable by Metadata Stripping)
Techniques such as Google SynthID embed microscopic statistical patterns directly into the visual pixel frequency space. Stripping metadata does not alter these pixel patterns.
3. Visual AI Classifiers (Independent of Metadata)
AI detection systems (e.g., Hive, Illuminarty) analyze the visual textures, unnatural skin reflections, and lighting anomalies of the pixels themselves, regardless of whether metadata exists.
4. Honest Privacy Guarantee
Online Metadata Tools guarantees genuine removal of all header-level metadata and C2PA manifests. We never make misleading claims about bypassing neural visual AI classifiers.
Frequently Asked Questions
C2PA (Coalition for Content Provenance and Authenticity) is an open technical standard backed by Adobe, Microsoft, Google, OpenAI, and the BBC. It embeds tamper-evident cryptographic JUMBF manifests (Content Credentials) into media containers, documenting who created or edited a file, which software or AI model was used, and the cryptographic signature chain.