CNN CLASSIFIER · GRAD-CAM EXPLAINABILITY

Is that photo real, or was it generated?

CIFAKE analyzes an image with a trained convolutional neural network and shows you exactly which regions influenced its decision — REAL or AI-generated — with a Grad-CAM visual explanation and a confidence score.

MODEL CNN · sigmoid INPUT MODEL-SIZED RGB STATUS READY
32×32
Input Resolution
2
Output Classes
Grad-CAM
Explainability
<1s
Inference Time

How the analysis works

Every upload goes through the same trained pipeline used in the CLI — now wrapped in a browser-friendly workflow.

1. Upload

Drag and drop or browse for a JPG, JPEG, or PNG image you want to verify.

2. Classify

The trained CNN scores the image and returns a REAL/FAKE label with a confidence percentage.

3. Explain

Grad-CAM highlights the pixels that most influenced the verdict, shown as a heatmap overlay.

Ready to analyze an image?

Get a REAL/FAKE verdict, confidence score, and Grad-CAM explanation in seconds.

Start Analysis