8 billion connected devices already carry a microphone: speakers, cameras, wearables, sensors, robots. They are always on, always present.
Yet almost none of them can hear danger. When a real emergency happens, no device understands the situation or raises the alert on its own.
Falls, cries for help, abnormal machine noises, early signs of distress at home, in care facilities, in public spaces or on the factory floor: critical situations are almost always announced by a sound.
Humans and animals hear these signals to stay safe. Machines, so far, stay deaf to them.
Buttons are not pressed.
Screens are not watched.
Alerts come too late.
Sonaid is embedded Sound AI. It gives any device with a microphone the ability to hear and recognize critical sound events in real-world, constrained environments, without voice commands, language or user interaction.
By analyzing acoustic signals rather than speech content, Sonaidās AI detects early warning sounds such as:
The technology is language-agnostic by design, privacy-preserving, and directly embeddable into existing connected devices equipped with microphones.
Sonaid addresses OEMs and device manufacturers who want to add hearing to their products, across six high-impact markets:
By giving connected devices the sense of hearing, Sonaid helps manufacturers detect events earlier, protect people better and build safer environments, anywhere sound exists.
<aside> šļø
Logos :
</aside>
<aside> š
SUMMARY
</aside>
<aside> š
</aside>
<aside> š„
Videos :
VideĢo preĢsentation techno Sonaid.mp4
</aside>
<aside> šø
Photos :
</aside>
Sonaid originates from cutting-edge French research in sound analysis.
The core technology was developed within Inria, in collaboration with leading academic and industrial partners, including Tokyo Metropolitan University and international research institutions.
From the outset, the research focused on a critical gap: how to detect ambiant sounds in real-world environments, sounds that traditional speech or language-based donāt understand.
During his doctoral research, Sonaidās founder, Nicolas Turpault, deliberately took a different path from conventional voice or speech recognition technologies.
Instead of analyzing what is said, Sonaid applies the same approach as Voice AI to any sound, focusing on what is happening in the environment.
Sonaidās AI is designed to recognize concerning acoustic signals ā such as falls, distress vocalizations or abnormal noises ā independently of language, accents or spoken content.
This language-agnostic approach makes the technology highly scalable, privacy-preserving and robust, even in noisy or constrained environments.
By late 2023, the technology reached industrial maturity, enabling its transition from academic research to real-world deployments and leading to the creation of Sonaid.
The solution is designed to be embedded directly into connected devices equipped with microphones, without relying on user interaction, voice commands or continuous screen monitoring.
Today, Sonaid addresses OEMs and device manufacturers who want to give their connected products the sense of hearing.
The technology enables manufacturers to unlock multiple high-impact applications, including:
<aside> <img src="/icons/calendar-day_blue.svg" alt="/icons/calendar-day_blue.svg" width="40px" />
Key milestones 2024: Minimum Viable Product launched in France 2025: Collaborations on home safety, work safety & senior care facilities usecases 2026: European launch with Arkea Care in Telecare Services
</aside>
Most telecare solutions still rely on buttons or wearables, devices that often fail when users cannot act. Sonaid enables a new generation of telecare through passive, always-on sound intelligence, embedded directly into connected devices to detect critical situations without buttons, cameras or language dependency.
Operating discreetly in the background, this approach allows telecare to be adopted earlier, without visible emergency devices, helping safety technologies blend naturally into everyday life. This model has already been validated through a strategic collaboration with ArkƩa Care.
Telecare still relies on buttons and wearables, devices that fail precisely when users cannot act or forget the wearable
<aside> š¢
</aside>
In senior living facilities, ensuring residentsā safety requires continuous vigilance, often relying on staff presence or intrusive monitoring systems. Yet cameras and wearables raise ethical, privacy and acceptance concerns, while staffing constraints limit round-the-clock supervision.
Sonaid brings a non-intrusive, sound-based approach to care facilities. Embedded into existing connected devices, its AI detects critical acoustic signals (falls, distress calls or abnormal sounds) without cameras, wearables or language dependency, enabling safer environments while respecting residentsā dignity.
This approach has already been explored in real-life care settings through a collaboration with Emeis, illustrating how sound intelligence can support care teams without adding complexity or surveillance.
<aside> š¢
</aside>
In industrial and corporate environments, many critical incidents occur outside direct supervision: falls, medical distress, aggression or abnormal situations, particularly for isolated or mobile workers. Traditional safety systems often rely on manual alerts or visual monitoring, which can fail when workers are unable to act or push a button when environments are not suited for cameras.
Sonaid enables a passive, sound-based layer of safety for the workplace. Embedded into existing connected devices, its AI detects early warning acoustic signals (falls, calls for help, abnormal noises) without buttons, wearables or cameras, and without any language dependency. This approach complements existing safety protocols without adding friction to workersā routines.