AI Image Metadata Checker
Drop an image on the box below to check if it has AI metadata. WipePic shows an “AI labels” group, a “Content Credentials (C2PA)” group and any prompt text it finds in the file. It does not detect pixel watermarks such as SynthID. Your image is not uploaded.
Drop your photos here
One photo or many photosOne photo, many photos or a folder (JPEG, PNG, WebP, HEIC, AVIF, DNG, GIF, TIFF)
How to check an image for AI metadata
- Drop your image on the box, or choose it from your device. You can also pick a folder.
- WipePic tells you what it found, for example “Found: AI labels, Content Credentials (C2PA)”.
- Click the image in the list. It opens and shows each field with its value, such as the AI label or the prompt text.
- To remove the data, click Remove hidden data and download the cleaned image. Your original file does not change.
Which AI marks are metadata
Some marks of AI live in the file, next to the pixels. The checker looks for these:
- Content Credentials (C2PA): a signed manifest that records where an image came from and how it was made or edited. WipePic shows it in the “Content Credentials (C2PA)” group.
- An AI label in XMP: the IPTC digital source type, for example
trainedAlgorithmicMediafor an image made by AI. WipePic shows it in the “AI labels” group. - Prompts and settings in PNG text chunks: tools such as Stable Diffusion front ends (Automatic1111) and ComfyUI save the prompt, the seed and the settings as text inside the PNG.
- Software names and edit history: the name of the tool that made or edited the image.
Which tools add which marks differs, and it can change. See which AI generators add metadata for details.
What the checker cannot do
- It cannot see pixel watermarks, such as Google SynthID. They are part of the image, not metadata. Read does removing metadata remove SynthID.
- It cannot tell that an image is AI-made when the file has no mark. Screenshots, re-saved copies and images from apps that strip metadata carry no AI data, even if AI made them.
- It does not judge the picture. It does not look at faces, style or errors in the image, and it does not guess.
- It does not read records on a server. Some services keep credentials outside the file. The checker reads only what is inside the file.
If you find AI metadata
Anyone who gets the file can read these marks, and platforms can read them too. A PNG with a prompt can share your prompt and settings. To remove them, use remove AI metadata, or remove Content Credentials if only the C2PA record is the problem. Removing metadata does not change your duty to say that an image is AI-made where a platform, a client or the law asks for it.
Your image stays on your device
WipePic reads and cleans your images in your browser. It does not upload them, and you do not need an account. The site uses Google Analytics for visit statistics, but your images, their file names and their metadata are not sent anywhere.
Questions
How do I check if an image has AI metadata?
Drop the image on the box above. WipePic reads it in your browser and lists the kinds of hidden data it found. If the image has an AI mark in its metadata, you see "AI labels" or "Content Credentials (C2PA)" in the list. Click the image in the list to see each field and its value.
Can the checker tell me if an image was made by AI?
Only when the image carries a mark in its metadata. If the checker finds an AI label, a Content Credentials record or prompt text, that is a strong sign. If it finds nothing, you learn only that the file has no such mark. Metadata is easy to remove, and many apps strip it, so a clean file does not prove that a human made the image.
Does it detect SynthID or other invisible watermarks?
No. A watermark such as Google SynthID is built into the pixels, and WipePic reads metadata, not pixels. A checker that only reads the file cannot see it. If an image has such a watermark, the AI metadata list can still be empty.
What is the difference between AI labels and Content Credentials?
An AI label is a short field, for example the XMP digital source type "trainedAlgorithmicMedia", that says an image was made by AI. Content Credentials (C2PA) are a signed record that is stored in the file. They tell where the image came from and how it was edited. WipePic shows them as two separate groups.
Why does a PNG show prompts and settings?
Many local image tools, such as Stable Diffusion front ends like Automatic1111 and ComfyUI, write the prompt and the settings into text chunks of the PNG file. The checker lists these as hidden text. Click the image to read the entries. Sharing the file shares the prompt too.
Is my image uploaded to check it?
No. WipePic reads your image inside your browser. The image, its file name and its metadata are not sent to a server. There is no account, and it is free.
Can I remove the AI metadata after I find it?
Yes. Click Remove hidden data and download the cleaned copy. WipePic removes AI labels, embedded C2PA manifests, prompts and other hidden data without changing the pixels. It does not remove pixel watermarks. Platform rules and laws on telling people that an image is AI-made still apply to you.