AI Image Detector — Free AI Image Checker
An AI image detector looks for signs that a picture was made or edited with AI. This free online one reads the file's labels first, and an AI model on your device estimates the rest.
- Free — no sign-up
- Checks for ChatGPT and Gemini watermarks
- Your images never leave your device
Loading AI image detector...
Want to thank us? Tell your friends about LeanImg!
How to Use the AI Image Detector
Add an image.
Drop a JPG, PNG, WebP, AVIF or HEIC file on the box, paste it with Ctrl+V or choose it. One image at a time.Read the file's labels.
Content Credentials, AI labels and hidden watermarks are checked first.Get the estimate.
If nothing in the file settles it, an AI model downloads once and estimates how likely the image is AI-generated.
Who This AI Image Detector Is For
Anyone on a dating app
Check a profile picture of someone you've met online. An app that re-saved the photo has usually stripped its labels, so expect the model's estimate here, and treat it as a hint: real photos re-sent through chat apps are the ones it most often scores high.
Buyers and renters
Check listing photos before you pay or book a viewing. A photo that carries its camera's Content Credentials shows “Taken with a camera” next to the estimate.
Journalists and anyone sharing news
Before you share a picture, see whether its Content Credentials or a hidden ChatGPT or Gemini watermark say AI. A screenshot of a post has lost the credentials, so find the original file if you can.
Artists and art buyers
See whether a picture's file says it was made with AI and which company signed it. When the file says nothing, you'll still get the estimate and a heatmap.
Popular AI Image Generators and What They Leave in a File
ChatGPT and Gemini sign their images with Content Credentials, and all 39 we saved straight from them carried a signature from a signer on the official trust list. Such a file names who signed it, OpenAI OpCo, LLC or Google LLC. Both also hide an invisible watermark in the pixels, and Gemini's is Google's SynthID. When the credentials are gone, this tool looks for each watermark's pattern, which it found in re-saved ChatGPT copies but in Gemini files only while they were untouched. Grok is different. It signs with a certificate of its own that's not on the trust list, so the Grok files we tested show 99%. Some Stable Diffusion apps save their generation settings inside the image, which also shows 99% without naming the app. A file with none of this, including a copy stripped of it, gets the model's estimate.
Frequently Asked Questions
How do I check if an image is AI-generated?
Run the original file through this detector. You'll see a percentage with the reasons behind it, then a heatmap of where the model reacted once the Deep scan completes on your device. A screenshot or a chat-app copy has lost the file's labels, so it leaves the detector less to go on.
Can AI image detectors be wrong?
Yes, in both directions. In our October 2026 test 10 of 69 real photos landed between 25% and 75%, and 3 of them scored 50% or more. Of 114 AI images, 4 got past both the file check and the model. Photos re-sent through chat apps are where it slips most often, so a percentage on its own shouldn't be used to accuse anyone of using AI.
Can this tool find SynthID watermarks?
Sometimes. Google hides an invisible SynthID watermark in images its AI makes, and ChatGPT images carry an invisible watermark too. This tool looks for the fixed pattern each one leaves. In our tests it found the pattern in all 19 original ChatGPT images and in 73 of 76 re-saved copies, while for Gemini it worked only on untouched files: 19 of 20 originals and 0 of 60 re-saved copies. Google's SynthID Detector and OpenAI's checker read the watermark itself, though each of them uploads your image to its company's servers.
Does my image get uploaded?
No. The file check, the watermark check and the AI model all run inside this browser tab, and the only network requests fetch the model and its helper files from this site. Open your browser's network tab during a check and you'll see nothing carrying your image away.
What does a low percentage mean?
Only that the model found little sign of AI in the image's structure. No detector can prove a photo is real, so even a low result leaves room for an AI image the model doesn't recognize.
Can it tell which AI made an image?
Only when the file says so. If its Content Credentials were signed by the company behind the tool, you'll see that company's name, such as OpenAI OpCo, LLC. Generator settings mark a file as AI-made without naming the app. We don't guess the generator from how the image looks, because the best-known detector in our tests named the wrong one for 3 of 8 AI images.
What are Content Credentials?
A signed label that some cameras and AI tools write into an image file, also known as C2PA. A valid signature from a signer on the official trust list proves which company signed the file and that it hasn't changed since. Re-saving, resizing and screenshots usually strip it, which is why a missing label tells you nothing either way about where the image came from.
Why do the reasons point in different directions?
Because the file and the pixels can disagree. A photo can carry camera data while the model still sees signs of AI in it, and we show both and change neither. Camera data is easy to edit, and detectors misjudge some real photos.
Can it show which parts of a photo were edited with AI?
Only when the file says so. Some phones and editing apps record an AI edit in the photo's Content Credentials, and then you'll see “AI-edited”. We tested browser models that mark edited areas, and the best one found about 9% of the edits while leaving 95% of untouched photos unmarked, so we don't mark edits. The heatmap answers a narrower question: it shows where the model reacted on a resized copy of the image. That doesn't prove those areas were made or edited with AI, and on a real photo the model gets wrong it glows just as brightly.
Why does the first estimate take longer?
The AI model (about 126 MB) downloads once and stays in your browser, so later checks start right away.
Can I check a video?
No. This tool checks still images, one at a time. For a video, take a screenshot of a frame and check that, though a screenshot loses the file's labels.
Is the AI image detector free?
Yes, with no sign-up and no daily limit. The checks run on your device, so there's no per-image server cost to pass on to you.
Can I use the result to prove someone used AI?
Not on its own. An estimate is a probability from a model that makes mistakes, so don't use it alone to accuse anyone. A Content Credential from a trusted signer is stronger evidence, because a known company signed it.
Does it work on screenshots and chat-app images?
Yes, with less to go on. Screenshots and chat-app copies lose the file's labels, so only the watermark check and the estimate are left. On copies of AI images resized and recompressed the way chat apps do it (1600 px, JPEG quality 70), the model caught 79 of 114.
How does an AI image detector work?
In two ways, and this one uses both. Some AI tools write evidence into the image: Content Credentials, an AI label, generator settings or an invisible watermark. Where there's none, an AI model looks for patterns AI leaves in the pixels and gives a probability. This detector reads the evidence first. Content Credentials from a signer on the official trust list that record AI show 100%. An AI label, generator settings or credentials from another signer that claim AI show 99%, and the ChatGPT or Gemini watermark pattern gives 97%. The model's estimate decides only when none of these turns up.
Does it work for checking real photos?
With limits, and it never calls a photo real. A photo signed by its camera with Content Credentials shows “Taken with a camera” next to the model's estimate, plus “No AI edits recorded” when the file lists none. Every photo still gets the estimate, and in our October 2026 test 3 of 69 original real photos came out at 50% or more. Check the original file, since a re-saved copy usually loses the camera's signature.
Can this AI image checker detect deepfakes and edited photos?
Partly. In our October 2026 test 3 of the 4 AI images the whole check missed were deepfakes. A photo is called AI-edited only when its Content Credentials or an AI label record the edit, and the tool doesn't mark which areas were changed.
Which AI image models can you detect?
There's no fixed list. The model reads patterns in the pixels and gives an estimate whatever made the image, but we don't publish a rate for each generator. On pixels alone it caught 16 of 19 Grok originals in our tests. Files from ChatGPT, Gemini and Grok can also settle it outright through their Content Credentials.
Which AI image detector is the best?
No single one, in our tests. In October 2026 we ran three detectors that check images on their own servers against this one, each on the files it scored. Every one of them missed some AI images, two of them missed fewer than this one on the same files, and one called fewer real photos AI. This one shows the reasons behind each percentage, so you'll see whether a label in the file or the model decided it.
Have more questions? Learn more about LeanImg and how we process your images.
Check if an image is AI-generated without uploading it
- Runs in your browser. The file check and the AI model run in this tab.
- No upload, no account. There's nothing to sign up for, and your files aren't sent to a server.
- Free to use.