Embed the world’s best voice processing technology into your solutions with Krisp SDKs
Krisp® AI-powered voice technology software is integrated into more than 100 million devices and has already improved over 10 billion minutes of voice communications
Krisp Technologies, Inc. is the leading developer of next-generation voice and audio technologies. Our award-winning software-based noise suppression technology is leveraging Deep Neural Networks to achieve an unprecedented level of noise immunity and voice quality.
Krisp has been deployed in over 100 million devices and has improved more than 10 billion minutes of voice communications.
Today, Krisp is serving companies around the world with our applications. Krisp is also powering Discord, the world’s leading gaming messenger with millions of its users utilizing voice daily.
Noise Suppression
Krisp leverages its Deep Neural Networks trained on tens of thousands of hours of audio to achieve an unprecedented level of quality of noise suppression.It performs equally well for both stationary and non-stationary noise types. Krisp noise suppression is language independent and is interoperable with all codecs.
Voice Activity Detection
Krisp’s Voice Activity Detection (VAD) technology detects if a given audio frame contains a human voice or not. Our VAD performs this function with extreme precision as it is based on the same underlying technology as our core noise suppression solution.
De-reverberation (aka Room Echo Suppression)
Krisp’s De-Reverberation technology, also known as Room Echo Suppression, delivers an unprecedented level of performance to enrich the audio quality of conference and collaboration rooms. Krisp leverages our Deep Neural Networks, trained on tens of thousands of hours of audio to detect reverberation and smoothly removes it from live audio while enhancing voice quality.
Noise Level Detection
Krisp’s Noise Level Detection (NLD) technology detects the Signal to Noise Ratio (SNR) level in the given audio frame. It’s used in situations when an app or device should detect the level of noise in the background (human voice is not counted as noise) and act on this level.
Supported Platforms
The SDK comes in three forms:
Complied C library
Compiled C library can run on Android, iOS, Windows and Mac. It can also be compiled for special-purpose embedded systems.
JavaScript library
Easily integrate into your web application. Requires WASM support (Chrome, Edge, Firefox, Safari).
Docker container
Krisp docker container can be hosted inside your server infrastructure and scaled to meet your needs.
Use Cases
Real-time communications
Voice and video conferencing, meeting room solutions, live-streaming, broadcasting, telephony.