Quick access

AI Image Checker

See whether an image says it was made or edited with AI. Reads Content Credentials (C2PA), IPTC AI labels, AI prompts saved in the file, and camera data. It reports only what the file declares, and the image never leaves your browser.

Use the original file

WhatsApp, Facebook, X, Telegram, and screenshots remove these labels. Download the image from where it was first posted, or ask the person for the file they saved.

What this check cannot see

Invisible watermarks such as Google SynthID live in the pixels and need their maker's own detector. Labels that were removed cannot be recovered. This tool does not guess from how an image looks, because pixel-based AI detectors are often wrong about real photos. Also check the source: who posted it first, and do other reliable sources show the same scene? Working with AI text instead? Clean up AI output.

About the AI Image Checker

Many AI image tools now label what they make. ChatGPT, Adobe Firefly, Google, and Microsoft sign their images with Content Credentials (C2PA), Meta and Midjourney write an IPTC "made with AI" tag, and Stable Diffusion and ComfyUI save the full prompt inside the PNG. This checker reads those labels and the camera data, verifies the signatures, and tells you exactly what the file declares and where it says so. It never guesses from the pixels.

How to use it

  1. 1 Drop an image here, choose one, or paste it from the clipboard. Use the original file, not a screenshot.
  2. 2 Read the verdict: AI-made, AI-edited, signed by a camera, or no label found.
  3. 3 Open Content Credentials, AI settings, and camera details to see each label and the field it came from.

Frequently asked questions

How can I tell if an image is AI generated?

Check the file for the labels AI tools add. Content Credentials (C2PA) from ChatGPT, Adobe Firefly, Google, and Microsoft record that an image was made with AI and are signed so they cannot be faked unnoticed. Meta and Midjourney add an IPTC "trained algorithmic media" tag, and Stable Diffusion saves the prompt in the PNG. If none are found, look for visible watermarks and odd details, and find where the image was first posted.

What are Content Credentials (C2PA)?

Content Credentials are a signed record inside a file that says who made it, with which app, and how it was edited. They follow the open C2PA standard, backed by Adobe, Google, Microsoft, OpenAI, Meta, Sony, Leica, and others. The signature breaks if the file is changed afterwards, so this checker shows whether the record still verifies and who signed it.

Why does it say no AI label was found on an AI image?

Labels are easy to lose. Most social networks and messengers, including WhatsApp, Facebook, X, and Telegram, strip metadata when you share an image, and screenshots and many editors drop it too. Some AI tools never add labels at all. So "no label found" only means the file does not declare anything, not that the image is real.

Can this detect SynthID or invisible watermarks?

No. Google SynthID and similar watermarks are hidden in the pixels and can only be read by their makers' own detectors. This checker reads the labels stored in the file's metadata, which it can show to you exactly.

Does it work with photos from my phone or camera?

Yes. It shows the camera make and model, lens, settings, date, and any GPS location stored in the photo. Phones such as the Google Pixel and Samsung Galaxy and cameras from Leica, Sony, and Nikon can also sign photos with Content Credentials, which shows up as "Signed as captured by a camera".

Is it an AI detector?

No, on purpose. AI detectors guess from the pixels and are often wrong, including about real photos. This tool only reports labels that are written in the file, so every result points to the exact field it came from.

Is my image uploaded?

No. The image is read in your browser and never leaves your device. The first time a file with Content Credentials is checked, the browser downloads the open-source C2PA reader (about 3 MB) from a CDN.