VoiceGenerationOptions

Defined in: packages/ai/src/types.ts:3070

Options for creating a voice.

Providers create voices in one of two ways, and some support both:

• design — synthesize a brand new voice from a text prompt.
• clone — derive a voice from referenceAudio of a real speaker.

At least one of prompt / referenceAudio is required; which ones an adapter accepts depends on the model. An adapter may require both — the only adapter today, elevenlabsVoiceDesign, always needs prompt and takes referenceAudio as an additional design reference.

Type Parameters#

TProviderOptions#

TProviderOptions extends object = object

Properties#

abortSignal?#

optional abortSignal?: AbortSignal;

Defined in: packages/ai/src/types.ts:3105

Effective abort signal composed by the activity from caller abortSignal and/or timeout. Adapters should forward this to the provider SDK when supported. Request-specific - never store on a global client config.

description?#

optional description?: string;

Defined in: packages/ai/src/types.ts:3091

Human-readable description stored alongside the voice

logger#

logger: InternalLogger;

Defined in: packages/ai/src/types.ts:3099

Internal logger threaded from the generateVoice() entry point. Adapters must call logger.request() before the SDK call and logger.errors() in catch blocks.

model#

model: string;

Defined in: packages/ai/src/types.ts:3074

The model to use for voice creation

modelOptions?#

optional modelOptions?: TProviderOptions;

Defined in: packages/ai/src/types.ts:3093

Model-specific options for voice creation

name?#

optional name?: string;

Defined in: packages/ai/src/types.ts:3089

Name to store the voice under in the provider's voice library. Providers differ on what this implies — ElevenLabs only persists a designed voice when a name is given. Read GeneratedVoice.saved to find out what actually happened.

prompt?#

optional prompt?: string;

Defined in: packages/ai/src/types.ts:3076

Text description of the voice to create, for design-capable models

referenceAudio?#

optional referenceAudio?: string | ArrayBuffer | File | Blob;

Defined in: packages/ai/src/types.ts:3082

Reference audio of the speaker to clone - base64 string, base64 data URL, File, Blob, or ArrayBuffer. For clone-capable models. Remote URLs are not accepted; read the file and pass the bytes.