📝 What is Krisp SDK?

Krisp SDK is a collection of real-time AI-powered technologies that improve speech clarity in real-time communication applications such as

• Online video and audio conferencing
• Online collaboration
• Streaming, broadcasting or podcasting
• Mobile calls

Currently, Krisp SDK has been deployed on more than 200M+ devices and is clearing 75B+ mins of speech every month. It is integrated inside apps such as Discord, RingCentral, and Twilio Video SDK.

Krisp SDK includes an impressive collection of AI-powered technologies that can be integrated into your application:

Real-Time AI Noise Cancellation

In real-time, separates background noises from a human voice, both inbound and outbound.

Real-Time AI De-Reverb

In real-time, separates reverb (room echo) from the human voice.

Real-Time Background Voice Cancellation

In real-time, separates background voices and noises from the main speaker's voice.

Real-Time AI Noise and Voice Statistics

In real-time, evaluates the level of noise and voice in the audio stream.

Session AI Noise and Voice Statistics

At the end of the session provides noise statistics based on the noise intensity levels and talk time

How Does it Work?

Runs on local device

• No server-side processing
• Integrates into the local application’s audio path (e.g. WebRTC)

Adds minimum latency

• Not noticeable for real-time communications
• Measured latency dependent on frame size and platform

Multiple models for different needs

• Supports NB, WB and Full-Band audio
• Supports different CPU speeds

Removed noise types

• Office, Street, Chatter, Baby, Animals, Keyboard, Fan, etc