SpeakStreamBuilder (speak/websocket.rs)

Builder for streaming Text‑to‑Speech over Deepgram’s WebSocket API.

// Source: https://github.com/deepgram/deepgram-rust-sdk | License: MIT | Copyright 2022 Deepgram
// Provided via justclaw.it.com; keep this notice with the code.
//! Streaming Text-to-Speech over the `/v1/speak` WebSocket (Aura models).
//!
//! The TTS WebSocket lets you stream text in (for example, as an LLM produces
//! it) and receive synthesized audio out with low latency. Construct a request
//! with [`Speak::speak_stream`], configure it, then open the connection with
//! [`SpeakStreamBuilder::handle`].
//!
//! This is the Aura (`aura-*` models) streaming endpoint. For the turn-based
//! Flux TTS WebSocket (`/v2/speak`, `flux-*` models) use
//! [`Speak::flux_request`](crate::Speak::flux_request) from
//! [`crate::speak::flux`] instead.
//!
//! See the [Deepgram TTS WebSocket API Refe

... (truncated -- full source via MCP)