Sanas

Sanas is a professional noise cancellation denoiser. Unlike the local RNNoise plugin, Sanas runs server-side over the Zero Runtime inference gateway, so it needs a gateway auth token. It occupies the pipeline’s denoise slot, cleaning the caller’s audio before the rest of the pipeline runs.

Setup

The denoiser connects to the inference gateway using your Zero Runtime auth token:

export ZERORUNTIME_AUTH_TOKEN=<token>

The token can also be passed explicitly to the factory.

Usage

Build the denoiser with SanasDenoise and pass it to the pipeline’s denoise slot.

from zeroruntime.inference import SanasDenoise

denoise = SanasDenoise(model_id="VI_G_NC3.0")

# Pipeline(denoise=denoise, ...)

import { SanasDenoise } from '@zeroruntime/js-sdk/inference';

const denoise = SanasDenoise({ model_id: 'VI_G_NC3.0' });

// Pipeline({ denoise, ... })

Parameters

Parameters for SanasDenoise.

Parameter	Type	Default	Description
`model_id`	`str`	`"VI_G_NC3.0"`	Sanas noise-cancellation model.
`sample_rate`	`int`	`16000`	Audio sample rate (Hz) the model expects.
`chunk_ms`	`int`	`20`	Audio chunk size in milliseconds.
`gateway_token`	`str`	`None`	Auth token for the inference gateway. Falls back to `ZERORUNTIME_AUTH_TOKEN`.
`base_url`	`str`	`None`	Override the inference gateway URL.

Import paths

SDK	Import	Constructor
Python	`from zeroruntime.inference import SanasDenoise`	`SanasDenoise(...)`
Node JS	`import { SanasDenoise } from '@zeroruntime/js-sdk/inference'`	`SanasDenoise({ ... })`

Noise cancellation runs first, so voice activity detection and speech-to-text see cleaner audio.

Plugins

Inference Gateway

Speech-to-text

LLM

Text-to-speech

Realtime models

Avatars

Turn detection

Voice activity detection

Noise cancellation

Setup

Usage

Parameters

Import paths

​Setup

​Usage

​Parameters

​Import paths

Setup

Usage

Parameters

Import paths