Media
1 item
Description
Analyses images on any page for AI generation signals: metadata, C2PA provenance, frequency watermarks, and statistical patterns. Lens analyses images on any webpage and tells you whether they were AI-generated - right where you're browsing. HOW IT WORKS When you visit a page, Lens quietly checks every image in the background. It looks for clues at several levels: • Where the image came from - some URLs and filenames are a dead giveaway that an image was AI-generated • Hidden metadata - AI tools often embed a digital signature at generation time, like a label sewn into a garment • The pixels themselves - AI-generated images have subtle patterns in their noise and structure that differ from real photographs • Invisible watermarks - some AI tools embed watermarks that survive even after editing Results appear as a small badge overlaid on each image. Hover the badge to see exactly what was detected and why. WHAT IT DETECTS Works across images from a wide range of AI image generators, including tools that embed C2PA metadata, spread-spectrum watermarks, or recognizable URL patterns. COMPLETELY PRIVATE • No images or data ever leave your browser • No account required • No telemetry, no analytics, no tracking • Image data is analyzed in memory and immediately discarded HONEST ABOUT LIMITATIONS AI image detection is a hard problem. Lens will not catch everything - images shared on social media often have metadata stripped, and some newer models produce pixel statistics that overlap with real photographs. When Lens flags an image it tells you its confidence level and the specific signals it found, so you can judge for yourself. Confidence levels: 🔴 AI Generated (≥90%) - strong evidence from verified metadata or watermarks 🟠 Likely AI (≥70%) - multiple corroborating signals 🟡 Possible AI (≥45%) - one or more weak signals 🟢 Probably Real (≥20%) - very low signal, likely a false alarm 🔵 No AI signals - nothing detected