Analyze image metadata for AI generator signatures. Educational visual detection guide included.
Drop an image here or click to upload
Supports JPEG, PNG, WebP
Even without metadata, trained eyes can spot AI images. Look for these common artifacts:
✋
Hands & Fingers
Extra or missing fingers, fused digits, unnatural bends
👁️
Eyes
Asymmetric pupils, reflections that don't match, glass glare inconsistencies
🔤
Text & Signs
Garbled letters, nonsense words, inconsistent fonts on signs
🌄
Background
Repeating patterns, melting objects, impossible geometry
💇
Hair
Strands that blend into background, unnatural flyaways
💍
Jewelry & Accessories
Earrings that don't match, floating items, merged accessories
💡
Lighting
Inconsistent shadow directions, multiple light sources on same subject
👂
Ears
Often distorted, cartilaginous details are wrong, asymmetric
Detecting AI-generated images requires a combination of metadata analysis and visual inspection. While AI generators have become remarkably good, they still leave telltale signs.
Many AI image generators embed metadata into images. Stable Diffusion often includes a "parameters" chunk in PNG files with the full prompt and model used. DALL-E and Midjourney may embed software tags. This tool reads those embedded signatures.
Many users strip metadata using tools like Photoshop or ExifTool before sharing. A clean metadata report does not mean the image is authentic — always combine metadata checks with visual inspection.