App for Monitoring the Patients with Suicidal Tendencies in Real Time
Psiho-vox is an AI-powered application currently in research and development, designed to monitor patients with suicidal tendencies between clinical visits. Built on clinically validated voice biomarker analysis, the system has achieved 81.6% classification accuracy in detecting varying levels of suicidal emotional intensity. It provides continuous, real-time insight into a patient’s mental state—supporting earlier and more informed clinical intervention. We are actively seeking clinical, research, and strategic partners to further validate, scale, and implement this technology in real-world healthcare settings.
1. Patient Module
AI virtual assistant that converses with the patient daily and analyzes the emotions specific to depression and suicidal tendencies embedded in their voice.
2. Doctor Module
The Doctor receives real-time information on patient's mental state, enabling timely clinical intervention when emotional intensity escalates.
Two Modules, One Mission
Research Partners
Institutul de Cercetari pentru Inteligenta Artificiala “Mihai Draganescu”
.jpeg)
Spitalul Clinic de Psihiatrie "Elisabeta Doamna"
Areus Technology
Universitatea de Medicina si Farmacie „Carol Davila”
.jpeg)
National Institute for Research and Development in Microtechnologies
Spitalul de Psihiatrie "Prof. Dr. Alexandru Obregia"
The Urgency We Cannot Ignore
Every year, over 727,000 people die by suicide worldwide. Alarmingly, up to 90% of these patients were previously assessed as low or no risk. For over 50 years, traditional self-report methods have failed to reliably predict suicidal behavior. Patients may consciously or unconsciously mask their true emotional state—leaving clinicians without accurate, timely data. The result: intervention often comes too late.
How does this work?
The Science Behind Psiho-vox
Voice biomarkers offer a theoretically grounded, non-self-report approach that captures emotional features patients cannot consciously suppress. Our CNN-based algorithm achieves 81.6% classification accuracy across three levels of suicidal emotional intensity, validated on 129 patients from two Romanian psychiatric hospitals.
Join Us in Advancing Suicide Prevention
Psiho-vox is built for the institutions that shape mental healthcare — hospitals, universities, and research centers committed to real outcomes. If you're ready to move beyond outdated tools, let's build the future together.
