Local forensic review

Check photos and videos for AI or deepfake signals.

Upload media, scan AI-generator metadata, sample visual evidence in the browser, and get a risk score with the exact signals that influenced it.

Made by Prisha Shinde under the supervision of Ranjit Sarangi in Aptech Lab.

Drop a photo or video here

PNG, JPG, WebP, MP4, MOV, and browser-supported formats work best.

Review history

Recent reports

No reports yet Analyze a file to create a local report summary.

Student project

How this app was made

This Deepfake Recognizer was made by student Prisha Shinde under the supervision of Ranjit Sarangi in Aptech Lab. It is built as a local-first browser project for learning, demonstration, and basic media review.

Frontend interface HTML sections create the upload area, result sidebar, reports, privacy note, and project panel.
Visual design CSS controls the white and gray layout, responsive panels, score ring, meters, and truth-check controls.
Media input Browser file APIs load photos and videos locally without uploading user files to a server.
Analysis logic JavaScript samples image pixels and video frames, checks metadata hints, and calculates a risk score.
Ground truth check The Real and AI-made buttons help students compare the detector result with a known sample label.
Deployment files Netlify settings and a build-check script keep the static app ready for presentation or hosting.

Privacy

Files stay in your browser.

This version does not upload media to a server. It uses browser APIs to inspect metadata, pixels, and video frames, then stores only short report summaries in this page while it is open.