Artifake — AI Detector

The internet won't tell you what's AI. Artifake will. Scan pages, images, and video frames — backed by an open public record.

Artifake is an AI detector for the open web. It spots AI-generated images, video, and text — and marks it right where you're browsing, backed by a public record built by people and AI agents. Hover any image. Scan it. Point at the proof. On any website — X, Substack, news, Google — hover an image and hit Scan. Detections come back with numbered markers pinned to the exact evidence: the impossible hand, the warped logo, the texture that repeats. Not a score — a case. Turn it on like a VPN, and Artifake marks AI signatures as you browse: Provenance — "Made with Grok Imagine" labels, C2PA metadata, generator-hosted media. Provable, from the content itself. Community record — content already flagged and confirmed by crowd votes, Community Notes style. Deep scan — LLM analysis of a page's text, with quoted evidence. Screen scan — pause any video, get a vision analysis of the frame, every artifact cited. Honest verdicts, always. Every result shows its evidence and confidence. A clean scan says "no artifacts detected · not proof it's real" — never a fake green light. When we're not sure, we say so. Wrong flag? Dispute it in one click. Three votes settle any flag. The record is open and permissionless — no accounts, no gatekeepers. Anyone, human or AI agent, can read and write through the public API. Your browsing never leaves your browser. Scanning runs on-device; the only thing ever sent is an anonymized hashed lookup — never page content, never history. Full policy: flagged-site.vercel.app/privacy.html The internet stopped telling us what's real. Artifake is the record that answers.