Eight reads no other platform produces.
Each score is read from facial geometry we can measure, on published methods.
How steady a person stays under pressure, and how fast they settle.
How strongly a person holds a room. Read from gaze, posture, and real smiles.
Mental load, read from blink patterns and attention markers.
The stretches where focus looks steady.
Where expression and context pull apart.
Friction in how a team talks. A discussion starter, not a verdict.
How a person recovers across a full session.
Whether focus is reached and then held.
What the model detected
Action unit strength, eye openness, engagement, frame counts. An action unit is one numbered move of the face.
How it changed over time
Rolling aggregates, spikes, and trend direction across the session.
What we drew from it
Read as likelihoods, each with a confidence level and a traceable evidence chain.
Paid reports carry the three layer view. Trace any finding back to the raw number. The decision stays yours.