The hearing layer for robotics
Overtone is building a hearing module for robots: hear, locate, and understand sound, even while moving. It runs on its own compute, not your robot's.
We're building Overtone with a small group of design partners.
Sound localization all around the robot, even in motion.
Added load on your robot's CPU and GPU.
Ears for every robot
Built to turn raw microphone input into events your robot can act on, all in one module.
Sound Localization. Built to pinpoint where every sound comes from in 3D space, even while your robot walks, drives, or turns, so it can head toward a voice or away from danger.
Runs on its own compute, not your robot's.
Hear every source. Follow the one that matters.
Overtone is built to separate overlapping sounds, tag each one with its direction, and let your robot decide what to track, ignore, or act on.
Hearing in motion
Built to hear while your robot walks, drives, or turns.
Real-time
Designed to process sound on the module, with no round trip to the cloud.
Its own compute
A closed module with its own processor, so your robot's CPU and GPU stay free for everything else.
Private by design
Built so raw audio stays on the module. Your robot only receives events and metadata.
From sound to action
Overtone is built to turn raw audio into structured signals your robot can understand and respond to.
Everything a robot needs to hear
- Noise suppression
- Filter out fans, machinery, and crowd noise so the sounds that matter come through clearly.
- Direction of arrival
- Know where a sound came from, with angle and distance for every source.
- Speaker separation
- Isolate individual voices when several people talk at once, and know who said what.
- Wake words & commands
- Recognize custom wake words and voice commands on the module.
- Anomaly detection
- Catch grinding bearings, leaks, and other mechanical faults by how they sound.
- Safety alerts
- Flag alarms, shouts, and breaking glass, even when nothing is in view.
Integrate seamlessly with any system
ROS 2 compatible. Python and C++ SDKs coming soon.
Pro-audio precision, built for robots
Over 20 years of professional audio experience, packed into a self-contained module that hears everything without taxing your robot.
From sound to signal
Sound localization
Separation between two talkers
Processed on the module
Native integration
Hears while walking or driving
* Working prototype stats
Shape how robots hear
We're building Overtone with a small group of design partners. Get early hardware, shape the roadmap, and work directly with our engineers.
Become a design partner