We collected and listened to:
Hours of Audio
Based on these, we learned and developed a neural network called krispNet DNN. We enhanced it by adding our “secret sauce”, and the result is magical audio processing that can recognize and remove any noise.
Krisp adds an additional layer between your physical microphone/speaker and conferencing apps, which doesn’t let any noise pass through.
All audio processing happens directly on your laptop. We don’t store or send any of your voice data to the servers.