The technical foundation
How the engine works under the hood. Long-form research papers covering methodology, threat catalogues, validation results, and the honest limits of current-generation detection.
Deepfake detection guide
10,000-word cornerstone reference — the technical foundation of how synthetic video is detected today, by us and by the industry.
Read paperVideo forensics basics
The fundamentals — metadata, compression chains, encoding signatures, frame analysis. The building blocks every detection method relies on.
Read paperAI manipulation techniques
A catalogue of how synthetic video is currently made — diffusion, GANs, face-swap, lip-sync — and the forensic fingerprints each technique leaves behind.
Read paperAuthenticity framework
Our assessment methodology — the verdict bands, the trust-score calibration, the evidence-fusion model, and why we believe in admitting uncertainty.
Read paperValidation study
Dataset results and accuracy benchmarks across known-AI, known-real, and adversarial corpora. The honest numbers — false-positive and false-negative rates included.
Read paper