How It Works¶

How Moveris detects real humans vs. spoofs—in plain terms and technical detail.

While deepfakes and AI-generated content have gotten remarkably good at mimicking human appearance, they can't fake biology.

Our API analyzes involuntary physiological signals that occur naturally when real humans interact with cameras—subtle reactions that happen below conscious awareness and can't be replicated by even the most sophisticated generative models.

Unlike traditional liveness detection that relies on challenge-response actions or AI model training, we've taken a fundamentally different approach rooted in psychophysiology. We measure what the body does automatically, not what it's told to do.

The Science (Simplified)¶

When you're alive and looking at a camera, your body is constantly generating signals—micro-expressions, subtle movements, physiological reactions that cascade through your system. These aren't things you can control or fake; they're the signature of a living, breathing human nervous system.

But it goes deeper than isolated signals. Real humans exhibit cognitive coherence—the natural, split-second coordination between what you're thinking, what you're seeing, and how your body responds. When you react to a question, your pupil dilation, facial muscle timing, and micro-movements all sync in patterns that reflect actual neural processing. Deepfakes can replicate individual elements, but they struggle to maintain this multi-layered coherence across time. The signals don't just need to exist; they need to make sense together.

Our technology reads these signals through standard webcams, requiring no special hardware. Within seconds, we can determine whether we're looking at a real person or a sophisticated fake.

Why This Gets More Effective, Not Less¶

Here's the counterintuitive part: as deepfakes improve, our approach becomes more valuable.

Traditional detection methods look for flaws—artifacts, inconsistencies, or statistical fingerprints left by AI generators. As generative models evolve, they learn to eliminate these tells. It's an arms race that favors the attacker.

We're not in that race. We're not looking for what's wrong with the fake—we're confirming what's present in the real.

AI can learn to add realistic-looking blinks or micro-movements, but it can't generate authentic biological coherence. There's no actual nervous system processing stimuli, no real pupils responding to light changes, no genuine cognitive load creating coordinated physiological patterns. These require actual neural tissue, actual consciousness, actual life.

As deepfakes reach visual perfection, the only reliable differentiator becomes: Is there a real biological system behind this? When pixels become indistinguishable, measuring the presence of life becomes the only moat that matters.

That's what Moveris does. And that's why we get stronger as the synthetic world gets more convincing.

Getting the Best Results¶

Embed in Natural User Flows¶

Our API works best when users are naturally engaged with their screen. Rather than treating liveness verification as a separate step, integrate it into existing user activities:

Camera Positioning Matters¶

• Face the camera directly rather than at an angle
• Ensure adequate lighting for better visibility
• Stay relatively still during the brief capture window
• Position at a natural distance (head and shoulders framing is ideal)

Integration Tips¶