orate/dist/elevenlabs.d.mts

The AI toolkit for speech.

import { TextToSpeechRequest, BodySpeechToTextV1SpeechToTextPost, BodySpeechToSpeechV1SpeechToSpeechVoiceIdPost, BodyAudioIsolationV1AudioIsolationPost } from 'elevenlabs/api';

type ElevenLabsModel = 'eleven_multilingual_v2' | 'eleven_flash_v2_5' | 'eleven_flash_v2' | 'eleven_turbo_v2' | 'eleven_turbo_v2_5' | 'eleven_multilingual_sts_v2' | 'eleven_english_sts_v2';
declare const voices: {
    alice: string;
    aria: string;
    bill: string;
    brian: string;
    callum: string;
    charlie: string;
    charlotte: string;
    chris: string;
    daniel: string;
    eric: string;
    george: string;
    jessica: string;
    laura: string;
    liam: string;
    lily: string;
    matilda: string;
    river: string;
    roger: string;
    sarah: string;
    will: string;
};
declare class ElevenLabs {
    private apiKey;
    constructor(apiKey?: string);
    private createProvider;
    /**
     * Creates a text-to-speech synthesis function using ElevenLabs
     * @param {ElevenLabsModel} model - The model ID to use for synthesis. Defaults to 'multilingual_v2'
     * @param {keyof typeof voices} voice - The voice ID to use for synthesis. Defaults to 'aria'
     * @param {Omit<TextToSpeechRequest, 'text' | 'model_id'>} options - Additional options for the synthesis
     * @returns {Function} Async function that takes text and returns synthesized audio
     */
    tts(model?: ElevenLabsModel, voice?: keyof typeof voices | (string & {}), options?: Omit<TextToSpeechRequest, 'text' | 'model_id'>): {
        generate: (prompt: string) => Promise<File>;
        stream: (prompt: string) => Promise<ReadableStream>;
    };
    /**
     * Creates a speech-to-text transcription function using ElevenLabs
     * @param {BodySpeechToTextV1SpeechToTextPost["model_id"]} model - The model to use for transcription. Defaults to 'scribe_v1'
     * @param {Omit<BodySpeechToTextV1SpeechToTextPost, 'model_id' | 'file'>} properties - Additional properties for the transcription request
     * @returns {Function} Async function that takes audio and returns transcribed text
     */
    stt(model?: 'scribe_v1', properties?: Omit<BodySpeechToTextV1SpeechToTextPost, 'model_id' | 'file'>): {
        generate: (audio: File) => Promise<string>;
    };
    /**
     * Creates a speech-to-speech conversion function using ElevenLabs
     * @param {BodySpeechToSpeechV1SpeechToSpeechVoiceIdPost['model_id']} model - The model ID to use for conversion. Defaults to 'eleven_multilingual_sts_v2'
     * @param {keyof typeof voices | (string & {})} voice - The voice ID to use for synthesis. Defaults to 'aria'
     * @param {Omit<BodySpeechToSpeechV1SpeechToSpeechVoiceIdPost, 'audio' | 'model_id'>} options - Additional options for the conversion
     * @returns {Function} Async function that takes audio and returns converted speech
     */
    sts(model?: BodySpeechToSpeechV1SpeechToSpeechVoiceIdPost['model_id'], voice?: keyof typeof voices | (string & {}), options?: Omit<BodySpeechToSpeechV1SpeechToSpeechVoiceIdPost, 'audio' | 'model_id'>): {
        generate: (audio: File) => Promise<File>;
        stream: (audio: File) => Promise<ReadableStream>;
    };
    /**
     * Creates a speech isolation function using ElevenLabs
     * @returns {Function} Async function that takes audio and returns converted speech
     */
    isl(options?: Omit<BodyAudioIsolationV1AudioIsolationPost, 'audio'>): {
        generate: (audio: File) => Promise<File>;
        stream: (audio: File) => Promise<ReadableStream>;
    };
}

export { ElevenLabs };