import type * as ElevenLabs from "../index";
/**
 * All fields are optional unless noted for the **first** message.
 *
 * **First message requirements**
 * - `voices`: non-empty array of voice IDs (maximum 10 for `eleven_v3`; exactly 1 for `eleven_v3_conversational`).
 * - Credentials if not supplied via `xi-api-key` / `Authorization` headers or `single_use_token` query parameter.
 *
 * **Subsequent messages**
 * - Do not resend `voices`, `pronunciation_dictionary_locators`, or credential fields.
 */
export interface TextToDialogueWebsocketClientMessage {
    /** Dialogue lines to append for synthesis. */
    inputs?: ElevenLabs.TextToDialogueWebsocketVoiceInput[];
    /** Force generation of buffered text without closing the connection. */
    flush?: boolean;
    /** Flush buffers, emit remaining audio, send `is_final`, and close the WebSocket. */
    closeSocket?: boolean;
    /** Resets the 20s inactivity timer; performs no synthesis. */
    keepAlive?: boolean;
    /** API key for the first message if not provided via the `xi-api-key` header. */
    xiApiKey?: string;
    /** Bearer token for the first message if not provided via the `Authorization` header. */
    authorization?: string;
    /** Single-use token for the first message if not provided via the `single_use_token` query parameter. */
    singleUseToken?: string;
    /** Voice IDs to load for the session (first message only, required on first message). */
    voices?: string[];
    /** Optional voice settings object; same shape as `voice_settings` on [Create dialogue](/docs/api-reference/text-to-dialogue/convert). */
    voiceSettings?: Record<string, unknown>;
    /** Optional pronunciation dictionaries (first message only). */
    pronunciationDictionaryLocators?: ElevenLabs.PronunciationDictionaryLocator[];
}