amazon-polly-streaming

Amazon Polly bidirectional streaming over HTTP/2 with SigV4: implements StartSpeechSynthesisStream, the Polly API that boto3 does not yet expose (only the AWS Java SDK has it).

Async generator interface: async for chunk in client.start_speech_synthesis_stream(...) yields audio bytes as they arrive from Polly, with no need to wait for the full audio to be generated server-side.
Real bidirectional: events sent to Polly and audio events received are framed in AWS event-stream over HTTP/2, with rolling chunk-signatures (the same scheme amazon-transcribe-streaming-sdk uses for Transcribe streaming).
Typed: PEP 561 py.typed marker; full pyright-strict surface for downstream consumers.

Full documentation on readthedocs.

Installation

pip install amazon-polly-streaming

Quick start

import asyncio
from amazon_polly_streaming import PollyStreamingClient

async def main() -> None:
    client = PollyStreamingClient(region="eu-central-1")
    audio = b""
    async for chunk in client.start_speech_synthesis_stream(
        text="hello world, how are you today",
        voice_id="Matthew",
        engine="generative",
        language_code="en-US",
        output_format="mp3",
        sample_rate="24000",
    ):
        audio += chunk
    with open("hello.mp3", "wb") as fh:
        fh.write(audio)

asyncio.run(main())

The voice_id must be a generative voice supporting bidirectional streaming. See the Amazon Polly Generative voices page for the current list and supported regions.

Requirements

Python 3.13+
AWS credentials in the default chain (env vars, profile, or IAM role) with polly:StartSpeechSynthesisStream permission
A region supporting Polly bidirectional streaming (us-east-1, us-west-2, eu-central-1, eu-west-2, ap-southeast-1, ca-central-1 as of 2026-05)

License

This package is released under the Apache License 2.0. See LICENSE and NOTICE for details.