Gaze Verification Guide

The Agentive Gaze Tensor (AGT) system classifies eye movement into three biometric tensors: Cognitive (COG), Emotional (EMO), and Environmental (ENV). These form a unique gaze signature for authentication.

AGT Tensors

The three biometric tensor zones that power JETT Auth. Each tensor captures a distinct dimension of human cognition through iris movement analysis, forming a 3-dimensional biometric signature that is unique to every individual.

Every page in the OPTX documentation is tagged with its primary AGT tensor — visible in the Map of Augments and the header legend.

COG — Cognitive Tensor

Color: Yellow (#eab308) | Gaze Zone: Upper field | Iris landmarks: Proprietary mapping

The Cognitive tensor captures analytical attention — the patterns your eyes make when processing structured information, reading, reasoning, and solving problems. COG activation correlates with:

• Sustained fixation on text or data (long dwell times)
• Systematic scanning patterns (left-to-right, top-to-bottom)
• Saccade precision during focused reading
• Pupil constriction indicating concentrated focus

In the OPTX documentation, COG-tagged pages are conceptual and architectural: protocol specs, system architecture, how-things-work explanations, and API references.

EMO — Emotional Tensor

Color: Red (#f43f5e) | Gaze Zone: Lower-left field | Iris landmarks: Proprietary mapping

The Emotional tensor captures affective processing — the involuntary eye movements associated with emotional engagement, social cognition, and identity recognition. EMO activation correlates with:

• Rapid micro-saccades during emotional processing
• Pupil dilation responding to arousal or surprise
• Gaze aversion patterns (looking away during discomfort)
• Face-tracking fixations during social engagement

In the OPTX documentation, EMO-tagged pages relate to identity, personality, agent behavior, and human-facing interfaces: JOE, gaze verification, and the DOJO.

ENV — Environmental Tensor

Color: Blue (#60a5fa) | Gaze Zone: Lower-right field | Iris landmarks: Proprietary mapping

The Environmental tensor captures spatial awareness — how your eyes map physical and digital space, navigate interfaces, and track movement. ENV activation correlates with:

• Peripheral scanning (wide-field sweeps)
• Depth-of-field shifts (near-far accommodation changes)
• Motion tracking saccades following moving targets
• Spatial mapping patterns (building mental models of layouts)

In the OPTX documentation, ENV-tagged pages cover infrastructure, networking, on-chain addresses, edge hardware, and cross-chain bridging — the spatial fabric of the system.

Tensor Weights

Authentication uses the weighted combination of all three tensors, not any single one. The weight vector [COG, EMO, ENV] is normalized to sum to 1.0 and feeds proprietary polynomial session binding.

Integration Example

import { useGazeClassifier } from "@jettoptics/gaze-sdk"

function GazeAuth() {
  const { tensor, confidence, weights } = useGazeClassifier({
    camera: true,      // Enable webcam capture
    fps: 60,           // Classification frequency
    minConfidence: 0.7 // Minimum confidence threshold
  })

  // tensor: "COG" | "EMO" | "ENV"
  // confidence: 0.0 - 1.0
  // weights: { COG: number, EMO: number, ENV: number }

  // Store gaze data for attestation
  const attestGaze = async () => {
    await fetch("/api/gaze/store", {
      method: "POST",
      headers: { "Content-Type": "application/json" },
      body: JSON.stringify({
        gaze_x: gazePosition.x,
        gaze_y: gazePosition.y,
        cog_value: weights.COG,
        emo_value: weights.EMO,
        env_value: weights.ENV,
        confidence: confidence
      })
    })
  }
}

Minimum gaze duration threshold required for CSTB attestation (configurable governance parameter).

JETT Auth

JETT Auth is the unified identity layer for both JettChat modes. Gaze verification produces an opaque proof that AARON validates — no raw biometric data is stored or transmitted.

Proof of you and your attention.