Skip to content
Model infoModelsSnapshot: 15 Oct 2026

Krisp / VIVA voice isolation

Voice Isolation 2.5

Primary-speaker isolation optimized to preserve speech that downstream transcription systems might otherwise omit.

Audio processing
availableClosed weightsAudio processing
Official documentation (opens in a new tab)Reviewed 13 Oct 2026
7 of 20 specification fields collected
Audio sample rates16 kHzSee specification notes below
Reported latency15 ms algorithmic latencySee specification notes below
Input billingNot collectedKrisp licensed SDK
Output billingNot collectedNo inferred price

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Context / input limit
Not collected
Maximum output
Not collected
Input modalities
AudioSource for Input modalities (opens in a new tab)
Output modalities
AudioSource for Output modalities (opens in a new tab)
Reasoning
Not collected
Tool calling
Not collected
Structured output
Not collected
Streaming
SupportedSource for Streaming (opens in a new tab)

Realtime frame or stream processing, not text-token streaming.

Batch processing
Not collected
Fine-tuning
Not collected
Knowledge cutoff
Not collected
Release date
Not collected
Parameters
Not collected
Architecture
Not collected
Languages
Not collected
Context caching
Not collected
Audio sample rates
16 kHzSource for Audio sample rates (opens in a new tab)
Provider-reported latency
15 ms algorithmic latencySource for Provider-reported latency (opens in a new tab)

Provider-reported algorithmic latency; excludes the rest of a call pipeline.

Deployment mode
CPU inference; no GPU requiredSource for Deployment mode (opens in a new tab)
Generation controls
enableLowPrecisionExecution uses native FP16 arithmetic where supported; otherwise keeps float32Source for Generation controls (opens in a new tab)

A runtime precision option, not a different model, cache policy or fine-tune.

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

Krisp licensed SDK

Provider source (opens in a new tab)
API model ID
Not collected

No exact public checkpoint selector confirmed for this access route.

Access
Request commercial SDK access; load the licensed model artifact and process audio frames on your infrastructure.
Regions
Customer-managed device or server deployment
Deployment mode
Licensed local inference; customer device/serverSource for Deployment mode (opens in a new tab)
Billing basis
Commercial SDK pricing by agreement; public developer tiers list no numeric model-level priceSource for Billing basis (opens in a new tab)

Meeting Assistant seat prices do not establish a developer inference tariff.

Endpoint pricing has not been collected. Check the provider for applicable rates.

Data & enterprise terms

SDK processing stays on the customer’s machines. The SDK security guide says Krisp does not collect audio or derived metadata and network access is for licensing. Optional metadata is generated locally. This policy does not describe the cloud Voice Translation API or meeting-recording service.

Read the service policy (opens in a new tab)

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

Krisp licensed SDK

Not verified

Loading model weights once or reusing an audio session is not a documented discounted prompt-cache feature.

How it works
Not collected

Loading model weights once or reusing an audio session is not a documented discounted prompt-cache feature.

Minimum reusable input
Not collected

Loading model weights once or reusing an audio session is not a documented discounted prompt-cache feature.

Lifetime & refresh
Not collected

Loading model weights once or reusing an audio session is not a documented discounted prompt-cache feature.

How to enable it
Not collected

Loading model weights once or reusing an audio session is not a documented discounted prompt-cache feature.

Usage fields
Not collected

Loading model weights once or reusing an audio session is not a documented discounted prompt-cache feature.

Billing rules
Not collected

Loading model weights once or reusing an audio session is not a documented discounted prompt-cache feature.

Limits & data boundaries
Not collected

Loading model weights once or reusing an audio session is not a documented discounted prompt-cache feature.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

What the measurements say

No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

Krisp identifies its AI models as proprietary, licenses the software and restricts extraction, modification and redistribution. The commercial SDK distributes licensed .kef inference artifacts through its customer portal, not open-source model releases. Historical/current generations are included only where named in provider sources.

Classification evidence (opens in a new tab)

Sources reviewed