Skip to main content
Image Checker

Real or faked?

Upload an image to detect deepfakes, AI generation, and digital manipulation. No signup required.

Drop an image here, or browse

JPEG, PNG, WebP, GIF Β· max 4 MB

What AI will detect

Frequently Asked Questions

What does 'deepfake' mean?
A deepfake is an AI-manipulated image or video where a person's face, body, or surroundings have been digitally altered using machine learning. Modern deepfakes can be extremely convincing β€” imperceptible to the naked eye β€” which is why automated pixel-level analysis is needed to detect them.
How accurate is AI image detection?
Our tool analyses pixel-level patterns, compression artifacts, and statistical anomalies associated with AI generation. While effective against current tools, no detector is perfect β€” highly sophisticated images or heavily post-processed photos may occasionally be missed. Use the confidence score as your guide: a low-confidence result should be treated as a rough indicator, not a conclusion.
Do you store my image?
No. Your image is never stored on our servers. We read the bytes in memory, compute a one-way SHA-256 fingerprint for caching the result, and discard the image data immediately. The hash cannot be reversed β€” the original image cannot be recovered from it.
What file types are supported?
We accept JPEG, PNG, WebP, and GIF files up to 4 MB. If your image is larger, try resizing or compressing it first β€” detection accuracy is not significantly affected by moderate compression.
Why was my image flagged as AI-generated?
Common triggers include overly smooth skin texture, unusual lighting consistency, subtle facial asymmetry from blending, or statistical patterns in the pixel data that differ from camera-captured images. If you believe the result is incorrect, you can try re-uploading a higher-quality version of the image or checking the confidence score β€” a low confidence value means the result is far from certain.
What if I disagree with the result?
AI detection is probabilistic β€” the result reflects the likelihood that an image was AI-generated, not a definitive verdict. The confidence score tells you how certain the system is. A low confidence score means the result should be treated as a rough guide. We recommend using it alongside other context: where the image came from, whether it was shared on a suspicious site, and whether metadata (EXIF data) looks plausible.
Is this free? What's the catch?
Completely free, no account required. Running detection calls costs us money, so if you find this useful, the Ko-fi link at the bottom of the result goes a long way. No ads, no signup, no catch.