Verifex — Forensic Sandbox

All-in-one AI deepfake
analyzer & detector

Verifex is your go-to, high-quality deepfake detector, making it easy to scan a picture and isolate synthetic artifacts with incredible accuracy.

But it doesn't stop there. With the forensic heatmap generator, you can instantly visualize hidden anomalies - exposing GAN and diffusion signatures seamlessly.

Core Technology

Our inference engine uses state-of-the-art neural networks to identify synthetic artifacts.

Xception CNN

Powered by the Verifex Engine, achieving 98.46% Validation Accuracy in detecting GAN and diffusion signatures.

Master Dataset

Trained on a massive collection of 227,000+ Master Images for peak forensic interpretation and reliability.

Secure Forensic Ledger

Cloud-integrated history tracking powered by Convex Ecosystem that allow history storage upto 10 pics.

The Origins of Verifex

Verifex is powered by its proprietary forensic engine, originally developed as a Final Year Project by the brilliant minds of DAE ICT (3rd Year) students at the Construction Technology Training Institute (CTTI), Islamabad.

Project Team

The architects behind the forensic neural engine.

Muhammad Bilal

Neural Network Optimization

Focused on fine-tuning the Xception CNN architecture and weight optimization.

Muhammad Sufyan

AI Inference & Signal Logic

Developed the signal processing pipeline and forensic inference logic.

Muhammad Uzair

Database Architecture & Sync

Engineered the Convex cloud integration and real-time data synchronization.

Umer Yasir

Full-Stack Development

Built the forensic dashboard, responsive UI, and frontend-backend bridge.

Uzair Ghani

Dataset Engineering

Managed the master dataset of 227,000 images and metadata labeling.

Mr. Hassan Ayaz

Project Supervisor

Lead Supervisor, ICT Department, CTTI Islamabad.

Future Roadmap

Expanding the frontiers of forensic detection.

Audio Forensics

Detecting AI-generated voice clones and synthetic speech patterns.

Browser Extension

A real-time scanning tool to verify images directly on social media platforms.

Video Forensics

Advanced frame-level analysis to detect temporal inconsistencies in deepfake videos.

Common Questions

Everything you need to know about Verifex's detection capabilities.

Verifex utilizes its proprietary forensic engine, powered by a fine-tuned Xception CNN. It has been validated on a master dataset of 227,000 images, achieving a 98.46% accuracy rate in detecting AI-generated manipulations and deepfakes.

Our architecture is "Integrated," combining a Python-based AI inference server with the Convex cloud ecosystem for real-time data synchronization. This ensures zero-latency processing and secure storage of forensic reports.

Verifex was developed as a Final Year Project by a team of DAE ICT (3rd Year) students at the Construction Technology Training Institute (CTTI), Islamabad, under the supervision of the ICT Department.

Yes. We prioritize privacy by using secure authentication via Clerk and encrypted cloud storage via Convex. Forensic analysis is performed in a secure sandbox, and history is only accessible to the authenticated user.

Currently, Verifex is optimized for high-fidelity image analysis. However, our 2026 roadmap includes "Temporal Inconsistency Analysis" for video deepfakes and "Synthetic Speech Pattern Recognition" for audio voice clones.

Expose the Synthetic Truth.

All-in-one AI deepfake analyzer & detector