Cardiopulmonary care,
modeled

Cardiopulmonary care,
modeled

Cardiopulmonary care,
modeled

Benchmarked against published comparators and clinically-accepted gold standards.

Ourmodelsanalyzereal-worldsignals,fromphysiologytobehavior,todeliverpersonalizedrecommendationsforpatientsandreal-timeinsightsforclinicians.

Ourmodelsanalyzereal-worldsignals,fromphysiologytobehavior,todeliverpersonalizedrecommendationsforpatientsandreal-timeinsightsforclinicians.

KAI plus 9 proprietary models.

KAI plus 9 proprietary models.

2 validated

7 in active validation

Flag Clinical Risks

Reconciles wearable streams, device readings and patient-reported signals into a clinician-ready summary

↳ Decompensation signature · 4-signal convergence

Escalate

Readable data

Reconciles wearable streams, device readings and patient-reported signals into a clinician-ready summary

↳ HR trend ↓ 8 bpm · activity ↑

Stable

Transparent Clinical Suggestions

Care-plan drafts with the signals and reasoning made visible, grounded in the latest guidelines.

Clinic visit ≤ 48h. Multi-signal pattern, weight still stable.

Approve

Transparent Clinical Suggestions

Care-plan drafts with the signals and reasoning made visible, grounded in the latest guidelines.

Clinic visit ≤ 48h. Multi-signal pattern, weight still stable.

Approve

The portfolio

KAI orchestrates a portfolio of proprietary models running today on Kento's devices. Performance numbers reference internal validation against published comparators and clinically-accepted benchmarks.

Deployed1 model

KAI

Deployed

The clinical orchestrator and personalization agent across the model portfolio.

Tasks & content
Patient Personalization
Alerts & Insights
Clinician Intelligence
Drafts & Updates
Care Plan Automation
Signal Routing
Model Orchestration

Live across active cardiopulmonary deployments
Validated2 models

KENTO | BREATHE

Validated

Ventilatory threshold detection (VT1, VT2). Parameter-free deterministic mathematical detector.

r = 0.955
n = 1,079
VT2 correlation with gold-standard CPET
Outperforms Zignoli (2020) and tracheal-sound methods

Kento

KENTO | QRS

Validated

Proprietary R-peak detector. The foundation under everything else in the platform.

Performance measured in progress (initial benchmark outperforms state-of-the-art) for a 2 leads or less ecg

Kento
In ValidationPredict family

KENTO | Predict

In Validation

A family of longitudinal prediction models trained on Kento's proprietary multi-source ECG-outcome corpus. Each model targets a distinct clinical decision.

KENTO | BCG

In Validation

Beat-to-beat cardiac mechanics from ballistocardiography fused with ECG, respiration, and motion. No echo, no cuff.

AUC validation in progress
Heart failure decompensation prediction
Extending Aydemir, Shandhi, Inan et al., IEEE TBME 2020

KENTO | Predict Cardiac Age

In Validation

Biological cardiac age from continuous ECG.

State-of-the-art MAE
On par with full 12-lead ECG-derived cardiac age

KENTO | Predict Vascular Age

In Validation

Vascular biological age from integrated sensor analysis.

AUROC in validation

KENTO | Predict Abnormalities

In Validation

Multi-class classifier across 37 ECG abnormalities.

AUROC > 0.94
Overall, across 37 abnormalities

KENTO | Predict Arrhythmia

In Validation

Ventricular arrhythmia, atrial fibrillation, atrial flutter, supraventricular arrhythmias.

Predicts optimal treatment trajectories for personalized care planning

KENTO | Predict Pathway

In Validation

Predicts which cardiopulmonary program transitions yield best outcomes, and identifies in-program behaviors that accelerate clinical improvement.

Predicts optimal treatment trajectories for personalized care planning

KENTO | Predict Lead12

In Validation

Full 12-lead ECG reconstructed for clinical-grade diagnostic use. Usable at home or in clinic.

r > 0.957
Precordial leads V1-V6 reconstruction accuracy
α Research4 active threads

Pulmonary artery pressure

α Research
Non-invasive estimation

α

Aortic valve health

α Research
S2/S1 balance analysis

α

Pressure-Volume loop estimation

α Research

α

No cuff blood pressure

α Research
<5 mmHg accuracy

α

See how Kento models support real patients.

See how Kento models support real patients.

See how Kento models support real patients.

The portfolio

KAI orchestrates a portfolio of proprietary models running today on Kento's devices. Performance numbers reference internal validation against published comparators and clinically-accepted benchmarks.

Deployed1 model

KAI

Deployed

The clinical orchestrator and personalization agent across the model portfolio.

Tasks & content
Patient Personalization
Alerts & Insights
Clinician Intelligence
Drafts & Updates
Care Plan Automation
Signal Routing
Model Orchestration

Live across active cardiopulmonary deployments
Validated2 models

KENTO | BREATHE

Validated

Ventilatory threshold detection (VT1, VT2). Parameter-free deterministic mathematical detector.

r = 0.955
n = 1,079
VT2 correlation with gold-standard CPET
Outperforms Zignoli (2020) and tracheal-sound methods

Kento

KENTO | QRS

Validated

Proprietary R-peak detector. The foundation under everything else in the platform.

Performance measured in progress (initial benchmark outperforms state-of-the-art) for a 2 leads or less ecg

Kento
In ValidationPredict family

KENTO | Predict

In Validation

A family of longitudinal prediction models trained on Kento's proprietary multi-source ECG-outcome corpus. Each model targets a distinct clinical decision.

KENTO | BCG

In Validation

Beat-to-beat cardiac mechanics from ballistocardiography fused with ECG, respiration, and motion. No echo, no cuff.

AUC validation in progress
Heart failure decompensation prediction
Extending Aydemir, Shandhi, Inan et al., IEEE TBME 2020

KENTO | Predict Cardiac Age

In Validation

Biological cardiac age from continuous ECG.

State-of-the-art MAE
On par with full 12-lead ECG-derived cardiac age

KENTO | Predict Vascular Age

In Validation

Vascular biological age from integrated sensor analysis.

the sensitivity at 95% specificity or above is >90%

KENTO | Predict Abnormalities

In Validation

Multi-class classifier across 37 ECG abnormalities.

AUROC > 0.94
Overall, across 37 abnormalities

KENTO | Predict Arrhythmia

In Validation

Ventricular arrhythmia, atrial fibrillation, atrial flutter, supraventricular arrhythmias.

Predicts optimal treatment trajectories for personalized care planning

KENTO | Predict Pathway

In Validation

Predicts which cardiopulmonary program transitions yield best outcomes, and identifies in-program behaviors that accelerate clinical improvement.

Predicts optimal treatment trajectories for personalized care planning

KENTO | Predict Lead12

In Validation

Full 12-lead ECG reconstructed for clinical-grade diagnostic use. Usable at home or in clinic.

r > 0.957
Precordial leads V1-V6 reconstruction accuracy
α Research4 active threads

Pulmonary artery pressure

α Research
Non-invasive estimation

α

Aortic valve health

α Research
S2/S1 balance analysis

α

Pressure-Volume loop estimation

α Research

α

No cuff blood pressure

α Research
<5 mmHg accuracy

α

The portfolio

KAI orchestrates a portfolio of proprietary models running today on Kento's devices. Performance numbers reference internal validation against published comparators and clinically-accepted benchmarks.

Deployed1 model

KAI

Deployed

The clinical orchestrator and personalization agent across the model portfolio.

Tasks & content
Patient Personalization
Alerts & Insights
Clinician Intelligence
Drafts & Updates
Care Plan Automation
Signal Routing
Model Orchestration

Live across active cardiopulmonary deployments
Validated2 models

KENTO | BREATHE

Validated

Ventilatory threshold detection (VT1, VT2). Parameter-free deterministic mathematical detector.

r = 0.955
n = 1,079
VT2 correlation with gold-standard CPET
Outperforms Zignoli (2020) and tracheal-sound methods

Kento

KENTO | QRS

Validated

Proprietary R-peak detector. The foundation under everything else in the platform.

Performance measured in progress (initial benchmark outperforms state-of-the-art) for a 2 leads or less ecg

Kento
In ValidationPredict family

KENTO | Predict

In Validation

A family of longitudinal prediction models trained on Kento's proprietary multi-source ECG-outcome corpus. Each model targets a distinct clinical decision.

KENTO | BCG

In Validation

Beat-to-beat cardiac mechanics from ballistocardiography fused with ECG, respiration, and motion. No echo, no cuff.

AUC validation in progress
Heart failure decompensation prediction
Extending Aydemir, Shandhi, Inan et al., IEEE TBME 2020

KENTO | Predict Cardiac Age

In Validation

Biological cardiac age from continuous ECG.

State-of-the-art MAE
On par with full 12-lead ECG-derived cardiac age

KENTO | Predict Vascular Age

In Validation

Vascular biological age from integrated sensor analysis.

the sensitivity at 95% specificity or above is >90%

KENTO | Predict Abnormalities

In Validation

Multi-class classifier across 37 ECG abnormalities.

AUROC > 0.94
Overall, across 37 abnormalities

KENTO | Predict Arrhythmia

In Validation

Ventricular arrhythmia, atrial fibrillation, atrial flutter, supraventricular arrhythmias.

Predicts optimal treatment trajectories for personalized care planning

KENTO | Predict Pathway

In Validation

Predicts which cardiopulmonary program transitions yield best outcomes, and identifies in-program behaviors that accelerate clinical improvement.

Predicts optimal treatment trajectories for personalized care planning

KENTO | Predict Lead12

In Validation

Full 12-lead ECG reconstructed for clinical-grade diagnostic use. Usable at home or in clinic.

r > 0.957
Precordial leads V1-V6 reconstruction accuracy
α Research4 active threads

Pulmonary artery pressure

α Research
Non-invasive estimation

α

Aortic valve health

α Research
S2/S1 balance analysis

α

Pressure-Volume loop estimation

α Research

α

No cuff blood pressure

α Research
<5 mmHg accuracy

α