- Home
- AI Voice Isolators
- Krisp Voice Isolation
Krisp Voice Isolation
Berkeley, California, United States
What it is
Krisp Voice Isolation is a real-time audio model in Krisp VIVA for voice AI agents. It separates a primary speaker from competing voices, background chatter, and room echo before speech-to-text processing. Krisp says Voice Isolation 2.5 runs on CPU with 15 ms latency and reduced word errors by 46% in its benchmark. The VIVA SDK also lists Turn Prediction, Interruption Prediction, Voice Activity Detection, and signal detectors for accent, gender, and synthetic speech. Deployment can run server-side on premises or in a customer’s cloud, ahead of any speech-to-text engine, and the SDK is described as compatible with any STT engine. Integrations include LiveKit, Pipecat, and raw WebRTC pipelines. SDKs are listed for Python, Node.js, Go, and iOS, using PCM S16LE at 16 kHz mono. Trial access requires describing a use case; pricing is on request.
For and against
For
- Separates voices and room echo
- 15 ms latency stated
- Runs on CPU
- Works with any STT engine
- Multiple SDKs and integrations
Against
- Pricing is available on request
- Audio format is PCM S16LE, 16 kHz mono
- Enterprise support requires the Enterprise plan
The verdict
Krisp Voice Isolation focuses on preprocessing audio for voice-agent systems, with additional prediction and detection models in the VIVA SDK. Its deployment options and integrations support varied server-side architectures, while access requires a use-case request.
Who it suits
Krisp Voice Isolation suits teams building voice AI agents that need audio cleanup before transcription. The Early stage plan is aimed at startups, early-stage builders, and teams experimenting with Voice AI.
Plans and prices
Early stage
VIVA and RTC models · apply now
Read 8 Oct 2026 · krisp.ai
Enterprise
Higher quality models · server-side batch processing · model customization · dedicated engineering support
Read 8 Oct 2026 · krisp.ai
On the record
- Additional models
- The VIVA SDK includes Turn Prediction, Interruption Prediction, Voice Activity Detection, and Signal Detectors for accent, gender, and synthetic speech.krisp.ai · read 8 Oct 2026
- Company
- Krisp's site identifies the company as Krisp Technologies, Inc. and lists its address in Berkeley, California.krisp.ai · read 8 Oct 2026
- Deployment
- VIVA runs server-side on-premises or in a customer's cloud and is designed to sit before the speech-to-text engine.krisp.ai · read 8 Oct 2026
- Integrations
- Krisp lists LiveKit, Pipecat, and raw WebRTC pipelines as supported integrations, and says the SDK works with any STT engine.krisp.ai · read 8 Oct 2026
- Performance
- Krisp says Voice Isolation 2.5 delivers 15 ms latency, runs on CPU, and reduced word errors by 46% in its benchmark.krisp.ai · read 8 Oct 2026
- Purpose
- Krisp VIVA provides real-time voice models that clean, understand, and transform audio for voice AI agents.krisp.ai · read 8 Oct 2026
- SDKs and audio format
- The VIVA page lists Python, Node.js, Go, and iOS SDKs and specifies PCM S16LE, 16 kHz mono audio in and out.krisp.ai · read 8 Oct 2026
- Security
- Krisp directs customers to its Trust Center for details about security practices, compliance standards, policies, controls, and security documentation.krisp.ai · read 8 Oct 2026
- Support
- The Enterprise developer plan includes dedicated engineering support and custom SLAs.krisp.ai · read 8 Oct 2026
- Supported platforms
- Krisp says its AI Voice SDK library is available for Windows, Mac, Linux, Web (JS/WASM), iOS, and Android.krisp.ai · read 8 Oct 2026
- Trial access
- Krisp says teams can describe their use case to be set up with the SDK and a trial; the page does not state a trial duration.krisp.ai · read 8 Oct 2026
- Voice isolation
- Voice Isolation separates the primary speaker from competing voices, background chatter, and room echo before speech-to-text processing.krisp.ai · read 8 Oct 2026
- Who it is for
- Krisp describes VIVA as voice infrastructure for voice AI agents and says its Early stage plan is aimed at startups, early-stage builders, and teams experimenting with Voice AI.krisp.ai · read 8 Oct 2026
