Skip to main content

Models and configuration

Models

EnumParametersApprox download
KittenModel.NanoInt815M28 MB including voices
KittenModel.Nano15M59 MB including voices
KittenModel.Micro40M44 MB including voices
KittenModel.Mini80M83 MB including voices

Voices

KittenVoice.Bella, Jasper, Luna, Bruno, Rosie, Hugo, Kiki, and Leo are available.

Configuration

const tts = await KittenTTS.create({
model: KittenModel.NanoInt8,
defaultVoice: KittenVoice.Luna,
speed: 1.05,
downloadRetries: 4,
ortNumThreads: 4,
maxTokensPerChunk: 400,
trimTrailingSilence: true,
silenceThreshold: 0.005,
maxSilenceTrimMs: 250,
});
OptionDefaultPurpose
storageDirectoryapp document directoryCache root
modelBaseURLHugging FaceSelf-hosted mirror directory
modelFilesnoneDirect local paths or bytes
downloadRetries4Total attempts per asset
ortNumThreads4ONNX intra-op threads
maxTokensPerChunk400Long-text inference limit
trimTrailingSilencetrueRemove near-silent chunk tails
silenceThreshold0.005Silence amplitude threshold
maxSilenceTrimMs250Maximum removal per chunk
phonemizerCECustom text-to-IPA implementation
forceRedownloadfalseRefresh assets during create
ortWasmPathpinned CDN on webSelf-hosted browser ORT files

Speed is clamped to the supported range of 0.5–2.0.