C2PA Detector: Check Images for Content Credentials
C2PA Content Credentials are the industry-standard "nutrition label" for digital media — a signed manifest in a file's header that records how it was made, including whether AI was involved. Gemini, ChatGPT, DALL·E, Grok and many camera and editing apps embed it, and Adobe, LinkedIn, TikTok and stock libraries read it automatically. This C2PA detector reads that manifest straight out of the file in your browser and tells you, instantly, whether an image carries Content Credentials — no upload, no account. If credentials are present and you need a clean file, the detector routes you to the remover that strips them.
What C2PA / Content Credentials Are
C2PA (the Coalition for Content Provenance and Authenticity) defines a tamper-evident manifest — often branded "Content Credentials" — embedded in JPEG, PNG and other files. It can name the generating model, timestamp the creation, and chain edits together. For AI images it's the clearest machine-readable proof of origin, which is exactly why so many platforms scan for it on upload. This detector looks for the C2PA, JUMBF and Content Credentials markers in the file bytes and reports what it finds.
Where C2PA Comes From
You'll find C2PA on images from Google Gemini, OpenAI's ChatGPT and DALL·E, xAI's Grok, Adobe Firefly and Photoshop exports, and a growing list of cameras and apps. Because it's metadata rather than a pixel signal, it's definitive to detect — present or absent — but also strippable by re-writing the file, unlike the SynthID pixel watermark some of these models add on top.
Detect, Then Remove the Credentials
When Content Credentials are found, the "Remove" button routes to the remover, which rewrites the file to drop the C2PA manifest and other provenance metadata while leaving the visible image untouched. If your image is from Gemini or ChatGPT, note that stripping C2PA alone won't remove SynthID — use the SynthID detector and remover for the pixel-level layer.