Voice technology that strips the room out of a call, so meetings sound like the person.
Noise, echo and voice separation for teams and individuals who work out loud all day.
Give every person on a call the same clarity a studio would give them.
Voice models good enough that distance stops shaping how people are judged.
We are 74 people across 9 countries, and hiring in audio research and product.
First noise-cancellation prototype ran on a laptop CPU.
Raised a $14.6M Series A to build the voice-separation model.
Named to a Top 50 AI list for applied audio research.
Cancelled 758 million minutes of background noise in one month.
2,140 businesses use it for clearer, shorter meetings.
Two engineers, one very loud café
Dariga Melkonyan and Petro Sahakian met over a call neither could hear. Six months later they had a signal model that separated the voice from the room — and no plan to stop there.