Psyche is building machine interoception: a state API for AI. We read the signals a person can't perform (heart rate variability, electrodermal activity, temperature, sleep) against each person's own baseline, and return their state to any system that acts on their behalf. How you are, not how you seem.

“I'm fine, just tired.” The voice is steady. But heart rate variability has been suppressed for thirty-six hours, electrodermal activity is elevated, and sleep has fragmented three nights running. Two channels, one moment, opposite stories. Psyche reads the divergence and says what it means.

Most systems that try to read people train on performances: actors, posed expressions, self-report. We work with the one channel that can't be performed. The signal is captured raw from the sensor, not filtered through consumer apps, and labeled against controlled induction rather than what people say they feel.

Models commoditize. No model gains a body. The paired dataset connecting signal to human state becomes more valuable as everything else gets smarter.

Psyche is born out of Carnegie Mellon's Human-Computer Interaction Institute. Our co-founder Mayank Goel directs the SMASH Lab there, with a decade of peer-reviewed work in passive sensing of human state. A recent pre-print predicts daily mood from consumer wearable signal alone, at roughly seventy percent accuracy. Our broader agenda is laid out in the thesis.

We are working with design partners and clinical collaborators now. If you believe the body is the signal AI is missing, book a conversation or write to us.