Streaming TTS

Connection Lifecycle

Understand the three phases of a streaming TTS WebSocket session.

Lifecycle phases

Phase	SDK Call	Description
Connect	`client.tts.stream()`	Opens a WebSocket connection and sends the text with config.
Receive	`async for audio in stream`	Yields binary audio chunks as they arrive from the server.
Done	`StopAsyncIteration`	The server signals completion and the async generator exits.

The SDK manages the WebSocket connection automatically inside AsyncShunyaClient.
Audio chunks are yielded as raw bytes in the format specified by response_format.
If the connection drops unexpectedly, the SDK raises a ConnectionError.