VERITY

Is This Image
AI Generated?

Upload the picture and get a plain-English answer — AI generated, edited, or a real photograph — along with the specific evidence behind it. Free, unlimited, no account.

Drop an image hereJPEG, PNG, WebP or HEIC · up to 18 MB

Checking video or audio instead? Use the full checker.

Six tells you can check by eye

Before any tool gives you an answer, these are the things a picture editor looks at. None of them is conclusive alone; two or three together usually settle it.

Hands, teeth and ears

Generators still lose count. Look for a sixth finger, a thumb on the wrong side, teeth that merge into one band, or an ear whose folds don't repeat on the other side of the head. Crop in tight — at full-image size these read as normal.

Text that falls apart

Signage, book spines, license plates and logos are where models break down fastest. Real lettering stays consistent in stroke width and spacing; generated lettering drifts into shapes that look like letters until you try to read them.

Light that disagrees with itself

Find the shadows and trace them back. In a real photo every shadow in a scene points away from the same source and softens at the same rate. Generated scenes often light the subject from one direction and the background from another.

Reflections that don't match

Mirrors, windows, sunglasses, glossy floors and eye highlights all owe a debt to the scene in front of them. When a reflection shows a different arrangement of objects — or a person's two eyes catch differently shaped highlights — the image was assembled, not captured.

Skin and fabric that are too clean

Camera sensors record uneven noise, heavier in the shadows than the highlights. Generated skin tends to be uniformly smooth, and repeated textures like knitwear, brickwork or crowd faces often tile subtly instead of varying.

Backgrounds that dissolve

Look at the far edges of the frame. Real optics blur backgrounds evenly with distance; generated backgrounds smear into approximate shapes — bystanders with no faces, buildings whose windows don't line up, street furniture fused together.

Sponsored

What the checker reads that your eyes can't

The camera's own record

Cameras write down make, model, lens, exposure and timestamp inside the file. Editors and image generators leave their own marks, and newer files can carry content credentials describing each step. A file claiming a camera that never shot that resolution has been through a pipeline nobody declared.

Compression and resave history

Every save stacks another compression layer onto a JPEG. When one region has been compressed a different number of times than its surroundings, something was pasted in. This is the single most reliable way to catch an edited real photo rather than a wholly generated one.

Noise structure

Sensor noise is uneven and predictable: stronger in shadows, tied to the sensor's colour filter layout. Generated images produce noise that is too uniform, or none at all in areas where a real exposure would have plenty.

When the answer will be uncertain

VTYAI says "not enough to tell" rather than guessing, and it is worth knowing when to expect that. Screenshots and downloads from social platforms have had their metadata stripped and their pixels recompressed, which erases most of the file-level evidence. Small crops leave too little area to measure noise across. Images that have been through heavy filtering, upscaling or "enhance" features carry generator-like traces even when the original was a genuine photograph. In all of these cases, chasing down the original file — from the person who took it, or via a reverse image search for an earlier posting — will tell you more than any detector.

Common questions

How can I tell if an image is AI generated?
Upload it above. VTYAI reads the file's hidden camera data, its compression and resave history, and its grain, lighting and texture patterns, then tells you whether it looks AI generated, edited, or authentic — and which specific findings led there. By eye, the fastest checks are hands and fingers, text on signs and labels, reflections that don't match their source, and backgrounds that dissolve into mush at the edges.
Is this AI image checker free?
Yes. Every check is free and unlimited, with no account and no sign-up. The site is supported by ads. VTYAI Pro exists only for people who want an ad-free experience.
Which images can I check?
Any JPEG, PNG, WebP, HEIC or similar picture file up to 18 MB. Screenshots work, but they lose the original camera data, so the answer leans more on visual analysis and the confidence will usually be lower.
Can an AI image detector be wrong?
Yes, and any tool claiming otherwise is overselling. Detectors flag statistical patterns, not proof. Heavily compressed, cropped or reposted images lose much of the evidence, and some generators now imitate camera noise deliberately. That is why VTYAI shows you every finding instead of a single score, so you can weigh the evidence yourself.
Does a screenshot from social media still work?
It works, but it is the hardest case. Uploading strips metadata and recompresses the picture, wiping the resave trail that makes tampering visible. Whenever you can, track down and check the original file instead.

Keep reading

How to spot an AI-generated photo without any special toolsNine practical checks — hands, text, reflections, skin texture, backgrounds, file data and more — that reveal whether a photo was generated by AI.How to tell if a video is a deepfakeA frame-by-frame checklist for face swaps, lip-sync fakes and fully generated clips, including what compression and re-uploads hide.What AI-detection tools actually measure — and where they failPlain-English explanation of metadata checks, compression forensics, texture statistics and model classifiers — plus the honest limits of each.How to tell if a voice message was cloned by AIVoice cloning now needs seconds of sample audio. Here is what a synthetic voice gets wrong — breathing, room sound, plosives, timing — and the one check that beats all of them.How to check a photo from breaking news before you share itDisaster and conflict photos spread fastest in the hours when they are least verifiable. A practical method for checking where an image really came from, using free tools.Spotting AI-generated photos in rentals, marketplaces and dating profilesGenerated images have moved into apartment listings, product pages and dating profiles. What to look for when money or a meeting is at stake, and the checks that settle it.Signal one: generation artefacts, and why models betray themselves in textureThe first of four signal types VTYAI examines. What a generation artefact actually is, where models leave them, and why they survive in some images and vanish in others.Signal two: physical consistency, or what light refuses to doThe second of four VTYAI signal types. How shadows, reflections, perspective and depth of field expose synthetic and composited scenes, and when physics checks mislead.Signal three: editing signatures, and how a real photo shows its scarsThe third of four VTYAI signal types. Compression history, noise discontinuity, resampling and clone traces reveal the edited real photograph — still the most common manipulation.Signal four: file metadata, the evidence that is easiest to read and easiest to forgeThe fourth of four VTYAI signal types. What EXIF, C2PA provenance and container data can establish about a file, why absence proves nothing, and how to read metadata without over-trusting it.

Share VTYAI

Help someone spot a fake — send them the link.