Models for Conversational AI
Krisp Audio SDK can be integrated into the audio pipeline on the server side to clean background noises and voices from audio streams. For example, putting Krisp NC before Voice Activity Detection (VAD) in the audio pipeline results in much better “turn detection” (aka “unwanted interruptions”) for Conversational AI.
Use Cases
- Conversational AI
- Speech-to-Speech models
- AI Voice Agents
Multiple Audio Stream Support
The SDK supports real-time processing of multiple audio streams using a single model loaded into memory, ensuring efficient memory utilization.
Supported Server Types
- Linux x86_64, armv8-a
- MacOS ARM, x86_64
- Windows x86_64
Supported Languages
Supported Frameworks
- KrispFilter is built into Pipecat
Recommended Models
Krisp offers different noise cancellation models optimized for different use cases, as shown in the following table.
- NC (noise cancellation) models remove background noises and background chatter
- BVC (background voice cancellation) models remove background noises and keep only primary speaker's voice
| Audio Source | Mic Requirement | NC | BVC | Krisp Model |
|---|---|---|---|---|
| Telephony, Cellular, Landline (8Khz) | Any | + | N/A | inbound NC (c7.n.s.9f4389) |
| Mobile, Desktop, Browser (WebRTC, 16kHz+) | Close-field talk or Headset | + | + | outbound BVC (c6.f.s.de56df) |
| Mobile, Desktop, Browser (WebRTC, 16kHz+) | Any | + | N/A | outbound NC (c8.f.s.026300) |
For more details on models refer to Model Guide.
Updated 5 days ago