Interface DeepgramBatchSTTProviderConfig

Configuration for the DeepgramBatchSTTProvider.

See

DeepgramBatchSTTProvider for usage examples

interface DeepgramBatchSTTProviderConfig {
    apiKey: string;
    model?: string;
    language?: string;
    mipOptOut?: boolean;
    fetchImpl?: {
        (input, init?): Promise<Response>;
        (input, init?): Promise<Response>;
    };
}

Properties

apiKey: string

Deepgram API key used for authentication. Sent as Authorization: Token <apiKey> in the request header.

model?: string

Deepgram model to use for transcription. nova-3 is Deepgram's general-purpose model. See https://developers.deepgram.com/docs/models-languages-overview for available models.

Default

'nova-3'
language?: string

BCP-47 language code, e.g. 'en-US', 'fr-FR', 'de-DE'. When omitted, Deepgram applies automatic language detection.

Default

'en-US' (set at transcribe-time if not configured here)
mipOptOut?: boolean

Opt requests out of Deepgram's Model Improvement Program by sending mip_opt_out=true: Deepgram keeps an opted-out request's data only for the duration necessary to process the request. Off by default: Deepgram's API reference asks callers to refer to its docs for pricing impacts before setting it.

Default

false
fetchImpl?: {
    (input, init?): Promise<Response>;
    (input, init?): Promise<Response>;
}

Custom fetch implementation for dependency injection in tests. When omitted, the global fetch is used. This allows tests to intercept HTTP calls without mocking globals.

Type declaration

    • (input, init?): Promise<Response>
    • Parameters

      • input: RequestInfo | URL
      • Optional init: RequestInit

      Returns Promise<Response>

    • (input, init?): Promise<Response>
    • Parameters

      • input: string | Request | URL
      • Optional init: RequestInit

      Returns Promise<Response>

Default

globalThis.fetch