Artifake

How Artifake works

Three ways to check anything you see. Every verdict shows its evidence, and clean scans never pretend to be proof.

Hover

Hover any image

A Scan badge appears on large images on any site. One click to analyze.

Right click

Right-click an image

Pick "Deep scan image with Artifake" from the menu. Same result, works everywhere.

Toolbar

Open the popup

Scan the page's images, its text, or a paused video frame with one button each.

1 Images

Scan an image, see the evidence pinned on it

Hover an image and click Scan. If AI artifacts are found, numbered markers land on the exact spots: the warped logo, the impossible reflection, the texture that repeats. Click a marker to read what it means. Not a score, a case.

1 2 3

Image analysis: AI-generated · 92%

1

Visor reflection implausibleThe gradient is too clean and does not match any light source in the scene

2

Hand anatomyFingers merge where they should separate, a common generation artifact

3

Texture repetitionThe same fabric pattern tiles across the suit instead of varying naturally

LLM analysis · model judgment, not proof
2 Video

Check a video: pause, then scan the frame

Video is just frames. Pause on the moment that looks off, open the Artifake popup, and hit Check a video. We analyze exactly what is on your screen, artifact by artifact.

LIVE
Breaking: markets react to overnight announcementUpdated 2 min ago · unverified clip circulating
1

Pause the video on the frame you want to check

Open the popup and tap Check a video

Get a verdict on the frame with markers on the evidence, like the lip-sync mismatch pinned at 1

3 Text

Scan a page's text, with quoted evidence

One tap on Scan this page's text runs an LLM analysis of the article, post, or newsletter you are reading. If AI writing patterns show up, you get the exact quotes, not a mystery percentage.

In today's fast-paced digital landscape, staying ahead means more than keeping up.

It's important to note that these tools do not just save time. They unlock seamless workflows and elevate every part of your process.

Moreover, the results speak for themselves.

LLM analysis: 84% AI-likely

"in today's fast-paced..." "it's important to note" buzzword cluster
LLM analysis · model judgment, not proof

When nothing is found, we say exactly that

A clean scan means we found no artifacts. It is not a certificate of authenticity, and Artifake will never give one. Good AI content can pass a scan, which is why every check also consults the community record: content flagged and confirmed by crowd votes that no single scan could settle.

No artifacts detected · not proof it's real