identifAI Models

In a world where AI-generated content is reshaping the very fabric of reality, identifAI stands as a guardian of truth. Our mission is to provide the tools needed to verify content authenticity, ensuring that the future of digital information remains transparent, trustworthy, and accountable. The way we handle AI-generated content will redefine the digital landscape and shape the future of truth itself.
Shield symbol

Cybersecurity Approach

identifAI approaches generative AI risks as a cybersecurity problem, not as a static ML classification task.

This has three concrete implications that materially differentiate identifAI from established players:

  • Threat-driven detection (not benchmark-driven detection)
  • Rapid, use-case-specific model adaptation
  • Scientific defensibility and methodological transparency
Illustration of a neural network

De-Generative Models

Our core technology, the "De-Generative Model", is an AI system designed not to create, but to critically analyse. Its function is to forensically examine digital media—images, video, and audio—to distinguish between authentic and synthetically generated content.
Each modality is processed by a specialised model. For video files, our system executes a comprehensive pipeline: the asset is separated into frames and audio, each is independently scrutinised, and a heuristic engine then produces a final, holistic assessment.

identifai MODEL FAMILIES
VISUAL MODEL
REVELIO

Detects fully AI-generated images and video. Built for KYC onboarding, content verification and fraud prevention. Processes millions of assets daily at sub-half-second speed.

The FACE version detects face-swapped selfies, synthetic avatars and camera injection attacks at point of submission. Built to protect identity verification and onboarding flows.

VOICE MODEL
ETHER

Detects cloned and synthetic voice in live calls and recordings. Sub-half-second. No perceptible latency. Built for telecoms, call centres and enterprise.

latest model cards
id. text
REVELIO FACE II
The latest deepfake detection model for faces by identifAI
99.5%
detection accuracy
93.8%
artificial precision
13.9%
performance boost
98.7%
human precision
id. text
REVELIO V PLUS
The latest deepfake detection for videos and images model by identifAI
99.8%
detection accuracy
99.9%
artificial recall
+7.4%
performance gain over previous model
+15
new generators
recognised
id. text
QUIET
The latest deepfake detection for audio model by identifAI
91.83%
detection accuracy
90.27%
artificial recall
+7.4%
performance gain over previous model
93,39%
human recall
See our latest models performance
Performance 
Discover our technology and let one of our experts guide you to learn how easy it can be to know the truth regarding online contents. Book a demo now!
Book a demo

talk to a human expert

Tell us about your business. We'll come back to you within one business day.

Thank you!
Your submission has been successfully sent to our team
Oops! Something went wrong while submitting the form.

No sales pitch. Just a conversation.

We stand for truth